How Common Are Non-Acute Coronary Syndrome

Te ara tika o te hauora hapori

Journal of the New Zealand Medical Association Vol 134 | No 1537 | 25 June 2021

Improving care, reducing the burden

Spatial, temporal and socioeconomic patterns of illicit drug use in New Zealand assessed using wastewater-based epidemiology timed to coincide with census

Social impacts and costs of schizophrenia: The competition is next door! Why a a national cohort study using New Zealand voluntary approach to tobacco retailer linked administrative data reduction will never work Te ara tika o te hauora hapori Publication Information published by the New Zealand Medical Association

NZMJ Editor NZMA Chair Professor Frank Frizelle Dr Alistair Humphrey

NZMJ Production Editor NZMA Communications Manager Richard Beer Simon Bradwell

Other enquiries to: NZMA To contribute to the NZMJ, first read: PO Box 156 www.nzma.org.nz/journal/contribute The Terrace Wellington 6140 © NZMA 2021 Phone: (04) 472 4741

To subscribe to the NZMJ, email [email protected]

Subscription to the New Zealand Medical Journal is free and automatic to NZMA members. Private subscription is available to institutions, to people who are not medical practitioners, and to medical practitioners who live outside New Zealand. Subscription rates are below. All access to the NZMJ is by login and password, but IP access is available to some subscribers. Read our Conditions of access for subscribers for further information www.nzma.org.nz/journal/subscribe/conditions-of-access If you are a member or a subscriber and have not yet received your login and password, or wish to receive email alerts, please email: [email protected] The NZMA also publishes the NZMJ Digest. This online magazine is sent out to members and subscribers six times a year and contains selected material from the NZMJ, along with all obituaries, summaries of all articles, and other NZMA and health sector news and information.

Subscription rates for 2021 New Zealand subscription rates Overseas subscription rates Individuals* $349 Individual $486 Institutions $604 Institutions $650 Individual article $33 Individual article $33

*NZ individual subscribers must not be doctors (access is via NZMA Membership) New Zealand rates include GST. No GST is included in international rates. Note, subscription for part of a year is available at pro rata rates. Please email [email protected] for more information. Individual articles are available for purchase by emailing [email protected]

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 2 www.nzma.org.nz/journal CONTENTS

EDITORIALS 56 8 Antipsychotic and sedative Improving care, medication use in long-term care reducing the burden facilities providing dementia care Wayne Miles Etuini Ma’u, Janine Burton, Elizabeth Fussell ARTICLES 66 11 Social impacts and costs of Spatial, temporal and schizophrenia: a national cohort socioeconomic patterns study using New Zealand linked of illicit drug use in New Zealand administrative data assessed using wastewater-based Sheree Gibb, Naomi Brewer, epidemiology timed to Nicholas Bowden coincide with census Mackay Price, Chris Wilkins, VIEWPOINTS Benjamin J Tscharke, Tom Baker, Jochen F Mueller, Sam Trowsdale 84 Cannabis Legalisation and Control 27 Bill: should doctors be concerned? Reducing the MRI outpatient Guna Kanniah, Shailesh Kumar waiting list through a capacity CLINICAL CORRESPONDENCE and demand time series improvement programme 91 Heera Bhullar, Bernadette County, Zoledronate-induced anterior Stuart Barnard, Anne Anderson, uveitis, scleritis and optic Mary E Seddon neuritis: a case report 36 Laura E Wolpert, Andrew R Watts Does research help to inform a 95 district health board’s purpose? Vildagliptin-induced A qualitative thematic analysis of bullous pemphigoid clinician researcher views Annabelle Yi Zhang, Paul Jarrett Lorraine Neave, Duncan Reid, Brian McKenna 98 Endogenous endophthalmitis 43 following CT colonography How common are non-acute Deepesh Mehta, Helen Long, Neil Aburn coronary syndrome (ACS) diagnoses in patients with LETTERS suspected ACS investigated with 100 coronary angiography in New The competition is next door! Zealand? (ANZACS-QI 58) Why a voluntary approach to Charles Yao-Cheng Hoa, tobacco retailer reduction Mildred Lee, Seif El-Jack, Peter Barr, Mark Simmonds, Gerry Devlin, will never work Richard Portch Philip D Adamson, Michael Williams, Andrew J Kerr

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 3 www.nzma.org.nz/journal CONTENTS

100 YEARS AGO 113 103 Proceedings for the 254th Rapture of the Urinary Bladder Otago Medical School Research 1921 Society Research Speaker Awards, Wednesday 24 June 2020 PROCEEDINGS 115 104 Abstracts for the 253rd Proceedings for the 256th Otago Medical School Research Otago Medical School Research Society Summer Student Sepeaker Society Masters/Honours Student Awards, Wednesday 22 April 2020 Speak Awards, Wednesday 4 November 2020 109 Abstracts for the 255th Otago Medical School Research Society PhD Student Speaker Awards, Wednesday 19 August 2020

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 4 www.nzma.org.nz/journal SUMMARIES

Spatial, temporal and socioeconomic patterns of illicit drug use in New Zealand assessed using wastewater-based epidemiology timed to coincide with census Mackay Price, Chris Wilkins, Benjamin J Tscharke, Tom Baker, Jochen F Mueller, Sam Trowsdale Wastewater samples were collected across seven sites in three regions of New Zealand to coincide with the 2018 census. This provided a way to quantify drug use and sociodemographics and assess the accuracy of drug survey techniques. The data show the expected regional, sub-regional and temporal patterns of illicit drug use, such as more MDMA consumed on the weekends than during the week. The key take home messages is that, in combination with complementary data like censuses, surveys and seizure information, wastewater analysis can help guide harm reduction policy for the wellbeing of New Zealand. Reducing the MRI outpatient waiting list through a capacity and demand time series improvement programme Heera Bhullar, Bernadette County, Stuart Barnard, Anne Anderson, Mary E Seddon A partnership between Ko Awatea and the radiology department at Counties Manukau District Health Board aimed to match the demand for magnetic resonance imaging (MRI) with the available capacity. Using quality improvement techniques, the team successfully and sustainably reduced the MRI waiting list over a two-year period, reducing the number of patients waiting from nearly 2,000 to just over 400. The innovative solutions to segment the waiting list and match capacity to demand may be instructive for other radiology departments, and other waiting list scenarios. Does research help to inform a district health board’s purpose? A qualitative thematic analysis of clinician researcher views Lorraine Neave, Duncan Reid, Brian McKenna This study explored, from a district health board (DHB) staff perspective, what the enablers and barriers to doing research were and whether the outcomes of research activity helped it deliver best care. It has provided insight into how organisational factors can negatively influence engagement in and with research and thereby impact the potential for research to inform the DHB’s overarching purpose. These factors are all modifiable influences. The themes offer a direction for the DHB’s future research strategy, where research activity is integrated into practice, where research evidence is discussed and where research achievements are celebrated. The findings from this study will likely resonate across New Zealand’s DHB landscape. How common are non-acute coronary syndrome (ACS) diagnoses in patients with suspected ACS investigated with coronary angiography in New Zealand? (ANZACS-QI 58) Charles Yao-Cheng Ho, Mildred Lee, Seif El-Jack, Peter Barr, Mark Simmonds, Gerry Devlin, Philip D Adamson, Michael Williams, Andrew J Kerr Falsely activated coronary catheter laboratories for acute ST elevation heart attacks are low in New Zealand. Non-ACS conditions receiving invasive coronary angiography varies between DHBs requires more investigation.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 5 www.nzma.org.nz/journal SUMMARIES

Antipsychotic and sedative medication use in long-term care facilities providing dementia care Etuini Ma’u, Janine Burton, Elizabeth Fussell This study aimed to quantify the use of antipsychotic and sedative medications in residents with dementia in long-term care facilities in the Waikato and identify factors associated with the prescription of these medications. Despite the increasing evidence around the relatively poor efficacy and increased risks associated with antipsychotic and sedative prescription in individuals with dementia, this study shows the use of these medications in dementia care and psychogeriatric facilities remains high, with 48.2% of residents in this study currently prescribed an antipsychotic and 22.1% prescribed a sedative. Furthermore, less than 20% of those prescribed an antipsychotic or a sedative had the most recent dose change occur within the 12 weeks recommended by guidelines, with over 30% of those prescribed an antipsychotic and 48% of those prescribed a sedative not having their dose adjusted for over a year. With clear evidence of the risks of antipsychotics in dementia, the proportion of residents prescribed an antipsychotic or sedative in this study, in conjunction with the prolonged duration of prescription, is cause for concern and needs addressing. Social impacts and costs of schizophrenia: a national cohort study using New Zealand linked administrative data Sheree Gibb, Naomi Brewer, Nicholas Bowden We used data for the whole New Zealand population to compare people diagnosed with schizophrenia to similar people without schizophrenia. Compared to people without schizophrenia, people with schizophrenia experienced more adverse outcomes, including: unemployment; lower income; contact with the criminal justice system; and hospitalisation (for both mental and physical health reasons). The government costs (money spent by the New Zealand government on people’s social, health and other services) were six times higher for people living with schizophrenia than for people without schizophrenia. The biggest cost differences were in healthcare, benefits and social housing. Cannabis Legalisation and Control Bill: should doctors be concerned? Guna Kanniah, Shailesh Kumar The regulatory and policy changes about legalising cannabis correlate with heightened acceptance, reduced perception of risks and an increase in cannabis use in both adults and adolescents, and controlled access to cannabis can be associated with better outcomes. The evidence for use in treating psychiatric illnesses is still lacking, and people who at a greater risk of experiencing adverse mental health outcomes will still be vulnerable under the proposed legislation. Regular cannabis use predicts an increased risk of schizophrenia. Cannabis is known to disproportionately harm people who have either a risk of mental illnesses or have existing mental illness. This significant at-risk population will not be protected by the proposed legislation in its current form. Data from other countries showing the beneficial effects of cannabis use in psychiatric populations are limited and conflicting, and potential harms in patients with psychotic and mood disorders have been increasingly documented. Further high-quality studies examining the effect of cannabis/cannabinoids on mental disorders are needed before any such policy changes.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 6 www.nzma.org.nz/journal SUMMARIES

Zoledronate-induced anterior uveitis, scleritis and optic neuritis: a case report Laura E Wolpert, Andrew R Watts Zoledronate, a bisphosphonate medication used to treat osteoporosis, can cause inflammation of different parts of the eye. Inflammation of the iris, known as anterior uveitis, happens in about 1% of people treated with bisphosphonates. This paper describes a case of a female patient who developed inflammation of the iris, the sclera (the white of the eye) and the optic nerve following treatment with zoledronate. The competition is next door! Why a voluntary approach to tobacco retailer reduction will never work Richard Portch This paper describes the distribution of tobacco retailers across Tamaki Mākaurau and compares them with community pharmacies. We used a new methodology for attributing the deprivation and demographics of neighbourhood surrounding each retailer. Our research identified significant differences in how tobacco retailers and community pharmacies are distributed across the region, including that tobacco retailers are more likely to be in residential areas than community pharmacies and are more likely to have another tobacco retailer in close proximity to their location.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 7 www.nzma.org.nz/journal editorial

Improving care, reducing the burden Wayne Miles

he conclusion from a meta-analysis detection of psychosis showed that early of studies of the economic burden of treatment for first psychosis had positive Tschizophrenia was that the enormous effects on clinical and functional outcomes.5 burden is suggestive of the inadequate pro- Reasons for delay include the stigma vision of healthcare services to patients with attached to any diagnosis of a mental 1 schizophrenia. Better resource allocation illness, and especially to that of schizo- and policy-orientated research is seen as a phrenia. It is also well recognised6 that the solution to this problem. The article in this disorder itself causes a lack of awareness 2 New Zealand Medical Journal provides an of the symptoms (anosognosia), such that excellent summary of current state of play an individual experiencing psychotic in New Zealand. It highlights the health, symptom will not have awareness of these. social and economic outcomes for those with We have a long way to go the shift public the disorder. attitude towards mental illness and to Recent exploration of what matters for create safe environments where the young people with schizophrenia highlighted people experiencing their first episode of a the need to feel safe, the opportunity for psychotic disorder (illness onset is typically employment, access to good healthcare late adolescence to early adulthood) can support and having meaningful social rela- present. They will often be reliant on family tionships.3 These factors were very similar or close friends to assist them recognise a to findings of a local (unpublished) review of problem and seek help for it; thus the point the long-term outcomes for people who have of care needs to be user friendly for these been through our anti-psychotic treatment support people also. trials. Though I do not have the data to support The age of the institution may have passed my clinical impression, it is likely that many but the messages emerging from work such presentations with schizophrenia are in as that of Sheree Gibb et al2 might be that crisis situations where there are concerns we still have a long way to go to deliver about the risk of the person with the healthcare, social and occupational supports disorder. They do not get seen in primary to people with schizophrenia that does not health settings, where one would ideally disadvantage them or stigmatise them. want an illness like this to be assessed and There are good guidelines for the treated. By the time they present there is treatment of schizophrenia.4 I suspect usually considerable disruption in their the problem is that we do not deliver lives; relationships are challenged or according to these best evidence-guided disrupted; usual occupation or education recommendations. and recreational pursuits have ceased. This loss of the usual social supports will The time between onset of symptoms and coincide with a time of immense personal beginning treatment remains disappoint- angst; the traumatic experience of identity ingly high. We are used to strong pushes for dissolution, boundary loss and the like the early identification of cancer because should not be underestimated. One only there is strong evidence that early detection has to consider what it must be like to have and appropriate evidence-based interven- some sinister voice that continues to berate tions significantly alter the disease course. you or threaten you; then, if you are brave We do not see such initiatives for psychosis. enough to mention it to others, you then get An elegant Scandinavian study which was told they cannot hear the voice—how do you set up to examine differences with true early assimilate that?

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 8 www.nzma.org.nz/journal editorial

For my goal of dedicated public health whole, might stimulate people in influence campaigns that target early identification to take the challenge to address the public of and rapid interventions for those at risk health importance of reducing stigma, of schizophrenia, it will be necessary to thereby enhancing the possibility of early address the pervasive public prejudice and access to best evidenced treatments. discrimination. This is no simple task and Why we want to treat early and maintain the available worldwide evidence of success good treatment is because it not only for anti-stigma campaigns for mental illness produces better outcomes regarding illness 7,8 is not reassuring. Strategies are generally experience and function, but because it described as contact with consumers, is also linked to arresting (and perhaps education campaigns and social activism improving) loss of cortical matter9 and and protest. Both contact approaches and decreasing the effect of schizophrenia education campaigns do have small to on reducing life expectancy.10 A German medium effect sizes. There is no study of study11 carefully reviewed the achievement social activism approaches. of remission (using well defined criteria12) I am hopeful that dissemination of the and suggests there is good reason for costs of schizophrenia, both to the people expectation of positive outcome from good who experience it and to the population as a treatment.

Competing interests: Nil. Author information: Wayne Miles: Clinical Director Research and Knowledge, Waitemata DHB; Clinical Associate Professor Department of Psychological Medicine, University of Auckland. Corresponding author: Assoc Prof Wayne Miles, Clinical Director Research and Knowledge, Waitemata DHB; Clinical Associate Professor Department of Psychological Medicine, University of Auckland [email protected] URL: www.nzma.org.nz/journal-articles/improving-care-reducing-the-burden

REFERENCES 1. Chong HY, Teoh SL, Wu Dark F, Humberstone V, Clinical Neurosciences. DB-C, Kotirum S,Chiou Jablensky A, Killackey 2006; 60: 531-7. C-F, Chaiyakunapruk N. E, et al. Royal Australian 7. Corrigan PW, Morris Global economic burden of and New Zealand College SB, Michaels PJ, Rafacz schizophrenia: a systematic of Psychiatrists clinical JD, Rusch N. Challeng- review. Neuropsychiatric practice guidelines for the ing the public stigma Diseases and Treatment. management of schizophre- of mental illness: a 2016; 12: 357-73. nia and related disorders. meta-analysis of outcome 2. Gibb S, Brewer N, Bowden Australian and New studies. Psychiatric N. Social impacts and Zealand Journal of Psychi- Services. 2012; 63: 963-73. atry. 2016; 50 (5); 410-72. costs of schizophrenia: a 8. Morgan AJ, Reavley NJ, national cohort study using 5. Larsen TK, Melle I, Ross A, Too LS, Jorm AF. New Zealand linked admin- Auestad B, Haahr U, Joa I, Intervetions to reduce istrative data. N Z Med J. Johannessen O, et al. Early stigma towards people 2021; 1537(134):66-83 detection of psychosis: with severe mental illness: 3. Lloyd H, Lloyd J, Fitzpatrick positive effects on 5-year Systematic review and R, Peters M. The role of life outcome. Psychological meta-analysis. Journal context and self-defined Medicine. 2011; 41: 1461-69. of Psychiatric Research. well-being in the outcomes 6. Pia L, Tamietto M. 2018; 103:120-33. that matter to people with Unawareness in 9. van Haren NEM, Hulshoff a disagnosis of schizophre- schizophrenia: Neuro- Pol HE, Schnack HG, nia. Health Expectations. psychological and Cahn W, Mandl RCW, 2017; 20: 1061-72. neuroanatomical find- Collins DL, et al.Focal 4. Galletly C, Castle D, ings. Psychiatry and grey matter changes in

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 9 www.nzma.org.nz/journal editorial

schizophrenia across the population-based study ralistic trial. Psychiatric course of the illness: a (FIN11 study). The Q. 2012; 83: 187-207. 5-year follow up study. Lancet. 2009; 374:620-7. 12. Andreasen NC, Carpenter Neuropsychopharmacol- 11. Schennach R, Riedal M, WT, Kane JM, Lasser RA, ogy. 2007; 32: 2057-66. Obermeir M, Jager, M, Marder SR, Weinberger DR, 10. Tiihonen J, Lonnqvist J, Schmauss M, Laux G et et al. Remission in schizo- Wahlbeck K, Klaukka T, al.Remission and recovery phrenia: propsed criteria Niskanen L, Tanskanen and their predictors in and rationale for consen- A et al. 11 Year follow up schizophrenia spectrum sus. American Journal of of mortality in patients disorders: Results from Psychiatry. 2005: 162: 441-9. with schizophrenia: a a 1 year follow-up natu-

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 10 www.nzma.org.nz/journal article

Spatial, temporal and socioeconomic patterns of illicit drug use in New Zealand assessed using wastewater-based epidemiology timed to coincide with the census Mackay Price, Chris Wilkins, Benjamin J Tscharke, Tom Baker, Jochen F Mueller, Sam Trowsdale

ABSTRACT AIMS: A discrete experiment in wastewater-based epidemiology (WBE) timed to coincide with the census was used to investigate the spatial, temporal and socioeconomic patterns of illicit drug consumption in Auckland, Bay of Plenty and Canterbury. METHODS: For seven consecutive days over census week (6 March 2018), wastewater was sampled from seven wastewater treatment plants and analysed for methamphetamine, cocaine (as benzoylecgonine) and 3,4-methylenedioxymethamphetamine (MDMA). Detailed sewer catchment maps were developed and, together with the data, were used to analyse drug consumption. RESULTS: Methamphetamine (mean 22.9 ± 9.9 doses/day/1000 people) was the most consumed drug, followed by MDMA (mean 1.7 ± 1.5 doses/day/1000 people) and cocaine (mean 0.5 ± 0.3 doses/day/1000 people). Methamphetamine consumption (and to a lesser extent MDMA) was high compared to that reported for Western nations, while cocaine consumption was extremely low. Cocaine and MDMA consumption were higher in cities compared to towns. In contrast, methamphetamine was typically higher in towns. Cocaine and MDMA were consumed more at weekends. Methamphetamine use was more consistent throughout the week. MDMA and cocaine were correlated with socioeconomic advantage, whereas methamphetamine was correlated with disadvantage. CONCLUSIONS: This paper contextualises illicit drug use in three New Zealand regions containing 18.3% of the national population and confirms the pervasiveness of methamphetamine consumption in New Zealand towns. This work demonstrates how WBE can be used to explore the socioeconomic dimensions of drug use when duly combined with other data sources like censuses.

ccurate and timely information about legal repercussions) and can un- about drug consumption is import- derrepresent certain demographics (eg, Aant for informing health and en- homeless, young, rural).2,3 Drug seizures are forcement policy. Such information is often often linked to police priorities and may obtained from population surveys, drug reflect increased enforcement, resources seizures and hospital and drug treatment or chance encounters rather than drug admissions, which have important limita- availability.4 Hospital and treatment centre tions.1 Surveys can suffer from self-report admissions can fail to capture recreational biases (due to social stigma and concerns drug use and can under-represent those

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 11 www.nzma.org.nz/journal article

users that do not seek help or experience isolated, rural communities.2 Anecdotal medical problems. evidence, however, suggests that metham- Wastewater-based epidemiology (WBE) phetamine use is proliferating in towns and 17–19 is an established complementary approach rural communities, and this is corrobo- to assess drug use that can support these rated by data on the location of clandestine 20 21 other sources of information. Through laboratories and drug availability. WBE is 22–24 the chemical analysis of drugs excreted gaining traction in New Zealand, and the into reticulated sewage systems, WBE recently commissioned National Wastewater provides quantitative measures of commu- Testing Programme (NWTP) has expanded nity-scale drug consumption that are not national understandings of illicit drug use 25 subject to self-report biases and aggregates by reporting findings at the regional level. from all dwellings connected to the sewer To complement this work, we present data network. This mitigates some of the issues from WBE that was specifically timed to of under-reporting common to population coincide with the census and report on the surveys.1 WBE can be conducted frequently, spatial, temporal (within-week) and socio- which enables assessments of short-term economic patterns of illicit drug use in three fluctuations in drug use. regions of New Zealand. WBE has been used to assess spatio-temporal patterns in drug consumption at Methods international, national and regional scales.5–8 Sites and catchment mapping In Europe, cocaine and MDMA consumption Wastewater samples were collected at is reportedly higher in cities compared to seven wastewater treatment plants (WWTP) smaller towns, whereas methamphetamine across three regions of New Zealand (Table consumption tends to be similar in both.5,8 1; Figure 1). The WWTPs were selected In Australia, drug use is generally higher to cover a range of population sizes and in regional communities than cities, except land uses (ie, cities and towns) and enable for cocaine.9 However, these urban–rural both inter- and intra-regional compar- patterns vary across regions,9 a reminder isons to be drawn. Site selection was also of the need to sample across a range of pragmatic, based on our contacts in the communities. wastewater industry. Collectively, these sites To better explain and understand these service 18.3% of the New Zealand popu- spatial patterns of drug use, attention has lation. Catchment maps were developed by recently turned to examine correlations superimposing geo-referenced sewer pipe between WBE and socioeconomic infor- information onto census statistical area mation. For example, significantly higher 1 geographies26 and trimmed to remove methamphetamine use was observed in properties not connected to the wastewater areas characterised by socioeconomic disad- assets. Populations for each site were calcu- vantage.10 A few studies have timed WBE to lated as the sum of the 2018 census night (de coincide with the census to better estimate facto) population for these trimmed areas. drug consumption and relate such patterns The census-night dataset accounts for all to demographics.10–13 individuals physically present and therefore New Zealand has several drug moni- includes domestic and international tourists. toring systems in place, including the New There are well-documented issues with the Zealand Health Survey.14 Previous drug-mon- 2018 census, including a lower than antic- itoring studies, like the Arrestee Drug Use ipated response rate.27 This was not ideal, Monitoring study (NZ-ADUM) and the Illicit but through sampling across census we Drug Monitoring System (IDMS), conducted were able to capture the most accurate and physical interviews of frequent drug users representative picture of the people present in New Zealand’s main cities, uninten- in our study sites at the time of sewer moni- tionally creating the impression that drug toring. This is an improvement on much of use is an urban phenomenon.15,16 This is the published WBE literature that typically perhaps unsurprising given the relative relies on population estimates derived from difficulty of recruiting drug-using popu- the design capacity of the WWTP and/or a lations for physical interviewing in more previous census, both of which may be years out of date.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 12 www.nzma.org.nz/journal article

Wastewater sampling and chemical metabolite were quantified using a cali- analysis bration curve of the ratio between the signal response for the unlabelled authentic drug Wastewater was sampled daily at each standard and deuterated analogue. Samples WWTP for seven consecutive days coinciding were analysed for methamphetamine, with the New Zealand census on 6 March cocaine (as benzoylecgonine) and 3,4-meth- 2018. Samples were 24-hour composites ylenedioxymethamphetamine (MDMA). collected using time-proportional sampling Benzoylecgonine was selected as the target of 100 mL of raw, screened influent every residue rather than cocaine because benzo- 15 minutes starting 6 am daily. At the end of ylecgonine is solely an excretory by-product each 24-hour period, samples were mixed of cocaine consumption (unlike the parent and reduced to a 1-litre volume. Mechanical drug), and concentrations therefore will be failure of the autosampler at site C1 was unaffected by the flushing of cocaine.1 The compensated for by collecting three 1-litre limit of detection (LOD) for all metabolites grab samples every eight hours (at 8 am, 4 was 33 ng/L. pm and 12 am). At the end of each 24-hour period, samples were frozen, and at the end Calculation of per capita of the week they were transported on ice consumption to the University of Auckland laboratory. Consumption was estimated using an During transit, two samples leaked (A3 established back-calculation formula.30 Saturday and B1 Wednesday) and were not Briefly, mass loads (mg/day) were esti- analysed. Within two days of arriving at the mated by multiplying concentrations (mg/ lab, samples were defrosted, filtered through mL) by daily wastewater volumes (mL). a 0.2 micrometre cellulose filter and acid- Mass loads were multiplied by a correction ified to pH 2 with 2 M hydrochloric acid. The factor that accounts for the metabolite’s processed samples were refrozen at -80 oC average excretion rate and molecular and transported on ice to the University of weight ratio between the metabolite Queensland where they were stored frozen and parent drug: 2.56, 3.00 and 4.44 for and analysed within two months. methamphetamine, cocaine and MDMA, Sample analysis followed a validated direct respectively.28,30,31 These back-calculation injection analytical method.28,29 Samples factors are based on an average excretion were defrosted and spiked with deuteri- rate. Although people metabolise drugs um-labelled chemical standards to correct at different rates, these differences will for instrument variability and matrix effects likely average-out at the population level. during analysis. Drug concentrations were It is not our intention to discuss the uncer- measured by direct injection using liquid tainties of excretion rates, as this has chromatography-tandem mass spectrometry been done elsewhere.1,32,33 However, very (LC-MS/MS). Concentrations of each drug/ briefly, the back-calculation formula does

Table 1: Study site information.

Site Region Population Size Community Type A1 Auckland 4,841 Town

A2 Auckland 54,547 City with rural areas

A3 Auckland 239,522 City

B1 Bay of Plenty 48,513 City

B2 Bay of Plenty 87,298 Town

B3 Bay of Plenty 66,856 City with rural areas

C1 Canterbury 374,364 City with rural areas

Although B2 has a large population, this site has a large wastewater catchment comprising of primarily suburban and rural landcovers and hence was classified as a town.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 13 www.nzma.org.nz/journal article

Figure 1: Regional study site locations: Auckland, Bay of Plenty and Canterbury.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 14 www.nzma.org.nz/journal article

not account for sewage leakage from pipes higher in B1 and B3 than both A3 and C1 or stormwater infiltration. We therefore (Table 2; Appendix Table 1). Methamphet- checked and confirmed that there were no amine consumption was also significantly wastewater overflow events. Additionally, higher in A1 than C1. Temporal patterns stormwater infiltration and subsequent of methamphetamine use were different dilution was likely negligible given the low between sites. For C1 and all Auckland rainfall recorded in the three regions for sites, methamphetamine consumption was the duration of the study. Consumption was relatively consistent throughout the week, normalised by population and converted with mean weekday (Monday–Friday) to doses/day/1000 people by dividing by a and weekend (Saturday–Sunday) differ- mean standard dose: 30, 100 and 100 mg ences ranging from 15 to 21% (Figure 2). In for methamphetamine, cocaine and MDMA, contrast, methamphetamine consumption respectively.30 Please keep in mind that the was higher on weekends for the Bay of exact quantity of drugs consumed (in mg) Plenty sites, particularly B1 (32% higher) in a single dose will vary based upon vari- and B2 (69% higher). Methamphetamine use ations in local drug purity and individual was significantly positively correlated with consumer preferences.1 Dosages simply socioeconomic disadvantage: r = .402, 95% facilitate comparisons across countries and CI [.118, .616], p = .005 (Figure 3). between individual drugs. Weekends were Cocaine grouped as Saturday and Sunday, except for Cocaine was only detected in 21% of MDMA, which included Monday as it has a samples, with a mean consumption of 0.5 relatively long excretion time.34 ± 0.3 doses/day/1000 people. The highest Socioeconomic dataset average cocaine consumption was observed The 2018 New Zealand Index of Multiple for A3, followed by A2, B1 and B2 (Table Deprivation (IMD) provided a descriptor of 2). Cocaine was not detected at the other relative socioeconomic status.35 The IMD sites. Cocaine was detected infrequently aggregates 29 indicators of socioeconomic throughout the week, with use restricted to deprivation under seven broad domains weekends and discrete weekdays. Cocaine (employment, income, crime, housing, health, exhibited a moderate, albeit insignificant education and access to services) for the 2018 negative correlation with socioeconomic period across 6181 data zones that together disadvantage: r = -.562, 95% CI [-.966, .192], p cover all of New Zealand. Each data zone has = .091 (Figure 3). a single rank score (from one to 6181), with MDMA higher scores representing higher levels of MDMA was observed in 74% of samples, disadvantage. Where data zones intersected with a mean consumption of 1.7 ± 1.5 doses/ mapped catchments, data-zone rank scores day/1000 people. MDMA consumption were population-weighted and averaged. was significantly higher in A3 and C3 Statistics than B3 (Table 2; Appendix Table 1). For The Shapiro–Wilk test was used to assess all sites, MDMA consumption was higher the normality of data groups. Kruskal– on weekends compared to weekdays. Wallis followed by Dunn–Bonferroni Notably, for A2, B1, B2 and B3, mean MDMA post-hoc testing was used to compare drug consumption was between 71 and 247% consumption between sites. To assess higher on weekends (Saturday–Monday) relationships between drug use and socioeco- than weekdays (Tuesday–Friday). MDMA nomic disadvantage, daily drug consumption was significantly negatively correlated with values were correlated with IMD rank scores socioeconomic disadvantage: r = -.614, 95% using Spearman’s correlation. CI [-.796, -.327], p < .001 (Figure 3). Results Discussion Methamphetamine Overview of consumption Methamphetamine was detected in 100% Methamphetamine was the most widely of samples, with a mean consumption of consumed illicit drug for which we 22.9 ± 9.9 doses/day/1000 people. Metham- tested, followed by MDMA and cocaine, phetamine consumption was significantly which corroborates other work in New

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 15 www.nzma.org.nz/journal article

Figure 2: Observed weekly methamphetamine consumption (doses/day/1000 people) for all sites in Auckland (A), Bay of Plenty (B) and Canterbury (C). Saturday and Wednesday are missing for A3 and B1, respectively.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 16 www.nzma.org.nz/journal article

Table 2: Observed drug consumption rates (doses/day/1000 people) for methamphetamine, cocaine and MDMA for each site.

Methamphetamine Cocaine MDMA Site Mean (SD) Range Mean (SD) Range Mean (SD) Range A1 23.6 (3.6) 20.4–31.3

A2 21.7 (1.9) 18.8–24.3 0.6

A3 16.6 (4.3) 10.6–21.5 1.1

B1 36.6 (12.1) 24.4–53.6 0.4 (0.3) 0.2–0.9 1.8 (1.0) 0.9–3.5

B2 20.9 (11.6) 10.6–45.6 0.2 (0.0)

B3 30.4 (3.7) 24.6–34.8

C1 11.5(1.6) 9.7–13.6

Standard deviation could not be estimated for A1 given that MDMA was only detected once at this site.

Figure 3: Relationship between deprivation rank scores and observed methamphetamine (A) (n = 47), cocaine (B) (n = 10) and MDMA (C) (n = 35) consumption (doses/day/1000 people) across all sites. Grey lines represent 95% confidence intervals.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 17 www.nzma.org.nz/journal article

Zealand.15,23,24 The data highlight the perva- with observations from recent wastewater siveness of methamphetamine in New studies.22,25 The high methamphetamine Zealand and supports its prioritisation consumption observed in the Bay of Plenty in public health and enforcement policy. also supports the high availability reported Unlike methamphetamine (which is both for this region.21 The relatively high use of imported and domestically manufactured MDMA in Canterbury, and both MDMA and in New Zealand), MDMA and cocaine are cocaine in Auckland, may simply reflect the almost entirely imported.16 They have low presence of nightclubs in cities.41 Cities have availability and low observed use, espe- larger populations too, which can sustain cially cocaine, likely due to New Zealand’s larger and more diverse drug markets.42 geographic isolation, tight border controls Cocaine and MDMA is almost entirely and relatively small market. New Zealand imported to New Zealand,15,21 so it makes Police and Customs are reportedly concerned sense that their consumption was higher about moves from international gangs to in the sites with ports (A3 and C1) than in establish a larger cocaine market in New site B3, which is landlocked and without an Zealand.36,37 This makes sense given cocaine’s international airport. high street price; however, it appears that The spatial patterns of methamphetamine this has yet to come into fruition in these paint a contrasting picture. Corroborating a regions, as evidenced in the WBE data. recent study,21 the data show that metham- Global comparison of phetamine consumption is greater in towns methamphetamine, cocaine and compared to cities. Methamphetamine labo- MDMA consumption ratories are often located in more isolated areas to avoid detection.20 Gangs have Mean methamphetamine consumption reportedly sought to expand methamphet- was considerably higher in New Zealand amine markets in towns.21 Local production, than reported in Europe (Figure 4), which socioeconomic decline, fewer police likely reflects the European preference personnel and a lack of market choice likely for amphetamine sulphate over metham- contribute to this pattern.19,21 Wastewater phetamine and the wider availability of studies in Australia and the United States other illicit drugs.38 Methamphetamine have also observed high methamphetamine consumption was lower than reported in the use in more rural communities.7,9,43 United States, Canada and Australia. These countries provide much larger markets However, our data show that metham- than New Zealand. It has been estimated phetamine is not only used in towns. There that over 90% of methamphetamine traf- was high consumption observed at the city ficked to Oceania is destined for Australia.39 site B1, which is adjacent to New Zealand’s In contrast, mean cocaine consumption in busiest port. Perhaps this area is an entry New Zealand ranked well below that of the point for internationally supplied meth- United States, Australia and Europe (Figure amphetamine and/or pseudoephedrine (a 4). This reflects the low volume and high chemical precursor for methamphetamine) price of cocaine in New Zealand (€221 per and, therefore, a subsequent hotspot for gram) compared to, say, the United States domestic methamphetamine manufacture. (56 € per gram) and European countries This is consistent with the large number like Belgium (€50 per gram) and Italy (€80 of clandestine laboratories seized in the 20 per gram).40 Mean MDMA consumption was region. It is important to understand relatively high in New Zealand compared to site-specific supply and demand dynamics many parts of Europe and North America for drug policy, and clearly WBE adds (Figure 4). Given the paucity of cocaine, it knowledge to support these efforts. is likely that MDMA is the substitute New Within-week patterns Zealand party drug.3 MDMA consumption was higher on Spatial patterns weekends compared to weekdays across all The regional pattern of relatively high sites. This weekend spike in MDMA accounts methamphetamine consumption in the for the large standard deviation across all Bay of Plenty, and MDMA (and cocaine) sites. Cocaine had a similar pattern, with in Auckland and Canterbury, is consistent consumption largely restricted to weekends

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 18 www.nzma.org.nz/journal article

Figure 4: Methamphetamine, cocaine and MDMA consumption rates (doses/day/1000 people) estimated for New Zealand (this study) and other countries. Number in brackets is the number of WWTPs. European and North American data are from SCORE, which is presented in the National Wastewater Drug Monitoring Program Report 7 alongside Australian data.9 All consumption rates were converted using the same excretion factors and dosages used in this study. European and North American data were collected in March 2018. Australian data were collected in December 2018.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 19 www.nzma.org.nz/journal article

and the odd discrete weekday. MDMA has more expensive drugs like cocaine), meth- a low dependence liability.44 Coupled to the amphetamine may be more accessible to relatively high street prices of MDMA and lower-income individuals.16 Additionally, cocaine, the data suggest that habitual use clandestine methamphetamine laboratories (ie, daily) of these drugs is uncommon. are often more concentrated in disadvan- 20 Within-week patterns of methamphet- taged communities, likely making it more amine use were variable. For C1 and all widely available in such areas. However, Auckland sites, methamphetamine was the correlation was not strong (r = .402), so consumed relatively consistently throughout the data highlight that methamphetamine is the week, supporting previous data for these consumed by people in advantaged commu- regions.22–24 Methamphetamine is highly nities too. addictive, so frequent and consistent user These trends are different to the 2007/2008 consumption is common.15 Interestingly, New Zealand Alcohol and Drug Use Survey methamphetamine consumption was higher (NZADUS), which found no relationship on weekends for the Bay of Plenty sites, between neighbourhood socioeconomic particularly B1 and B2. This perhaps reflects status and the use of methamphetamine, the high number of bars and nightclubs in cocaine or MDMA.45 It could be that the B1. Other studies have found city popula- NZADUS simply did not capture heavy user tions to have more pronounced weekend groups.2 Perhaps people were reluctant to drug use compared to towns.6 B1 is also a participate in the survey for fear of stig- popular tourist destination, and the high matisation or legal repercussions. Another weekend consumption may be attributable, contributing factor is that drug surveys may in part, to tourists. not representatively capture rural popu- 3 It was surprising that different patterns lations. Regardless of the reason, our data of methamphetamine use were observed at highlight the complementary nature of WBE sites B1, A3 and C1, given that these are all to other measures of drug use. The next city sites (with bars and nightclubs). MDMA challenge is to best align the datasets as a consumption rates were higher in A3 and step towards transdisciplinary public health C1 compared to B1 (Table 2). It may be that policy and drug intervention. MDMA is more available and thus preferred as a recreational weekend drug in these Conclusion sites. This would suggest that the low avail- Wastewater-based epidemiology was ability of MDMA and/or high availability timed to coincide with the 2018 New Zealand of methamphetamine may facilitate recre- census. The data confirm the pervasiveness ational methamphetamine consumption. of methamphetamine in New Zealand This is concerning from a public health and supports its ongoing prioritisation perspective given methamphetamine’s rela- in public health and enforcement policy. 16 tively high dependence liability. As only There were inter- and intra-regional differ- a single week of samples were collected, ences in drug consumption. Cocaine and longer-term sampling is being undertaken to MDMA consumption were higher in cities, confirm these findings. whereas methamphetamine was generally Socioeconomic patterns higher in towns. Notably, high metham- To better explore these spatial patterns, phetamine consumption was observed at drug use was compared with socioeconomic Bay of Plenty’s urban site, highlighting the disadvantage. MDMA was significantly importance of site-specific supply dynamics negatively correlated with disadvantage. and local consumer preferences. Cocaine Cocaine was similar but not significant, and MDMA were consumed infrequently being influenced by the small number of throughout the week, with consumption detections. Conversely, methamphetamine largely restricted to weekends. Metham- was positively correlated with disad- phetamine was consumed more consistently vantage, which is consistent with local throughout the week in Canterbury and survey data.16 Due to methamphetamine’s all Auckland sites. In contrast, metham- relatively low street price (compared to phetamine consumption was higher on

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 20 www.nzma.org.nz/journal article

weekends in the Bay of Plenty. The data amphetamine was positively correlated. The show that methamphetamine is consumed research shows that WBE provides valuable both habitually and recreationally. Such data on the spatial, temporal and socio- patterns, however, cannot be accounted for economic patterns of drug use, and with by differences in urbanisation or population complementary information can be used to size alone. Expectedly, MDMA (and cocaine) help guide the development of nested local, consumption was negatively correlated with regional and national-scale drug policy in socioeconomic disadvantage, whereas meth- response.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 21 www.nzma.org.nz/journal article

Appendix

Appendix Table 1: Dunn–Bonferroni comparisons of methamphetamine consumption rates (doses/ day/1000 people) between sites. Bolded values are significant.

Site comparison Mean rank difference Std. error Adjusted p-value C1–A3 7.6 7.6 0.999

C1–B2 13.4 7.3 0.999

C1A2 18.3 7.3 0.265

C1–A1 22.3 7.3 0.05

C1–B3 33.4 7.3 < .001

C1–B1 34.8 7.6 < .001 A3–B2 -5.8 7.6 0.999

A3–A2 10.7 7.6 0.999

A3–A1 14.7 7.6 0.308

A3–B3 -25.8 7.6 0.015

A3–B1 -27.2 7.9 0.013 B2–A2 4.9 7.3 0.999

B2–A1 8.9 7.3 0.999

B2–B3 -20 7.3 0.133

B2–B1 21.3 7.6 0.108

A2–A1 4.0 7.3 0.999

A2–B3 -15.1 7.3 0.815

A2–B1 -16.5 7.6 0.646

A1–B3 -11.1 7.3 0.999

A1–B1 -12.5 7.6 0.999

B3–B1 1.3 7.6 0.999

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 22 www.nzma.org.nz/journal article

Appendix Table 2: Dunn-Bonferroni comparisons of MDMA consumption rates (doses/day/1000 people) between sites. Bolded values are significant.

Site comparison Mean rank difference Std. error Adjusted p-value B3–B2 8.0 5.3 0.999

B3–A2 13.3 5.3 0.188

B3–B1 15.5 5.5 0.076

B3–A3 20.1 6.2 0.019

B3–C1 -21.2 6.9 0.031 B2–A2 5.3 5.3 0.999

B2–B1 7.5 5.5 0.999

B2–A3 12.1 6.2 0.786

B2–C1 -13.2 6.9 0.824

A2–B1 -2.2 5.5 0.999

A2–A3 -6.8 6.2 0.999

A2–C1 -7.9 6.9 0.999

B1–A3 4.6 6.4 0.999

B1–C1 -5.7 7.0 0.999

A3–C1 -1.1 7.6 0.999

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 23 www.nzma.org.nz/journal article

Competing interests: Nil. Acknowledgements: We would like to thank Watercare Services Limited, Christchurch City Council, Ngāi Te Rangi and Tauranga City Council Safe Communities for their assistance with the sample collection and the facilitation of the research. We thank Arnuad Besse-Ciller for his help visualising the data. We are also grateful to the two anonymous reviewers and the section editor for their time to read and comment on the paper. Author information: Mackay Price: School of Environment, University of Auckland, Auckland. Chris Wilkins: Drug Research Team Leader, SHORE & Whariki Research Centre, College of Health, Massey University. Benjamin J Tscharke: Postdoctoral Research Fellow, Queensland Alliance for Environmental Health Sciences (QAEHS), The University of Queensland, Australia. Tom Baker: Senior Lecturer, School of the Environment, University of Auckland, Auckland. Jochen F Mueller: Group Leader, Queensland Alliance for Environmental Health Sciences (QAEHS), The University of Queensland, Australia. Sam Trowsdale: Senior Lecturer, School of the Environment, University of Auckland, Auckland. Corresponding author: Mackay Price, School of Environment, University of Auckland, Auckland [email protected] URL: www.nzma.org.nz/journal-articles/spatial-temporal-and-socioeconomic-patterns-of-illicit-drug-use-in-new-zealand-assessed-using-wastewater-based-epidemiology-timed-to-coincide-with-the-census-2

REFERENCES 1. van Nuijs A, Castiglioni 5. EMCDDA. Wastewater Intelligence Commission. S, Tarcomnicu I, et al. analysis and drugs: a National wastewater drug Illicit drug consumption European multi-city study. monitoring Report 7, March estimations derived Lisbon: EMCDDA, 2020. 2017. Canberra: Australian from wastewater anal- 6. Krizman I, Senta I, Ahel M, Criminal Intelligence ysis: A critical review. Terzic S. Wastewater-based Commission, 2019. Sci Total Environ. assessment of regional and 10. Choi PM, Tscharke B, 2011;409(19):3564-77. temporal consumption Samanipour S, et al. 2. McFadden Consultancy. patterns of illicit drugs Social, demographic, and The New Zealand Drug and therapeutic opioids economic correlates of food Harm Index 2016 (second in Croatia. Sci Total and chemical consumption edition). Wellington: Environ. 2016;566:454-62. measured by wastewa- Ministry of Health, 2016. 7. Lai FY, O’Brien J, Bruno ter-based epidemiology. 3. Prichard J, Lai FY, O’Brien R, et al. Spatial variations Proc Natl Acad Sci U S A. J, et al. ‘Ice Rushes’, Data in the consumption of 2019;116(43):21864-73. Shadows and Methylam- illicit stimulant drugs 11. Hart OE, Halden RU. phetamine Use in Rural across Australia: A Simulated 2017 nation- Towns: Wastewater nationwide application wide sampling at 13,940 Analysis. Current Issues of wastewater-based major U.S. sewage in Criminal Justice epidemiology. Sci Total treatment plants to assess 2018; 29(3):195-208. Environ. 2016;568:810-8. seasonal population bias 4. Willis K, Anderson J, 8. Ort C, van Nuijs AL, Berset in wastewater-based Homel P. Measuring the JD, et al. Spatial differences epidemiology. Sci Total effectiveness of drug law and temporal changes in Environ. 2020;727:138406. enforcement. Trends illicit drug use in Europe 12. O’Brien JW, Thai PK, & Issues in crime and quantified by wastewater Eaglesham G, et al. A criminal justice no. 406. analysis. Addiction. model to estimate the Canberra: Australian Insti- 2014;109(8):1338-52. population contributing tute of Criminology, 2011. 9. Australian Criminal to the wastewater using

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 24 www.nzma.org.nz/journal article

samples collected on Policing. 2018;12(2):231-42. assessment. Water Res. census day. Environ Sci 21. Wilkins C, Romeo 2011;45(15):4437-48. Technol. 2014;48(1):517-25. JS, Rychert M, et al. 29. Lai FY, Anuj S, Bruno 13. Tscharke BJ, O’Brien JW, Determinants of high R, et al. Systematic Ort C, et al. Harnessing availability of methamphet- and day-to-day effects the Power of the Census: amine, cannabis, LSD and of chemical-derived Characterizing Waste- ecstasy in New Zealand: population estimates water Treatment Plant Are drug dealers promoting on wastewater-based Catchment Populations for methamphetamine rather drug epidemiology. Wastewater-Based Epidemi- than cannabis? Int J Drug Environ Sci Technol. ology. Environ Sci Technol. Policy. 2018:61, 15-22. 2015;49(2):999-1008. 2019;53(17):10303-11. 22. Chappell A, Ashmore, 30. Zuccato E, Chiabrando C, 14. Ministry of Health. E. Wastewater Pilot Castiglioni S, et al. Estimat- Annual Data Explorer Programme. Institute of ing community drug abuse 2018/19: New Zealand Environmental Science by wastewater analysis. Health Survey. Wellington: and Research, New Environ Health Perspect. Ministry of Health, 2019. Zealand Police, 2019. 2008;116(8):1027-32. 15. Wilkins C, Prasad J, Barnes 23. Lai FY, Wilkins C, Thai P, 31. Khan U, Nicell JA. Refined HM, et al. New Zealand Mueller JF. An exploratory sewer epidemiology mass Arrestee Drug Use Monitor- wastewater analysis study balances and their appli- ing (NZ-ADUM) 2010–2016. of drug use in Auckland, cation to heroin, cocaine Auckland: SHORE & New Zealand. Drug Alcohol and ecstasy. Environ Int. Whariki Research Centre, Rev. 2017;36(5):597-601. 2011:37(7):1236-52. Massey University, 2017. 24. Wilkins C, Lai FY, O’Brien 32. Castiglioni S, Bijlsma L, 16. Wilkins C, Prasad J, Romeo J, et al. Comparing meth- Covaci A, et al. Evaluation JS, Rychert M. Recent amphetamine, MDMA, of uncertainties associated trends in illegal drug use in cocaine, codeine and with the determination New Zealand, 2006–2015: methadone use between of community drug use Findings from Illicit Drug the Auckland region and through the measurement Monitoring System (IDMS). four Australian states using of sewage drug biomark- Auckland: SHORE & wastewater-based epide- ers. Environ Sci Technol. Whariki Research Centre, miology (WBE). N Z Med 2013;47(3):1452-60 Massey University, 2017. J. 2018;131(1478):12-20. 33. Gracia-Lor E, Zuccato E, 17. Bayer K. Meth plague 25. New Zealand Police. Castiglioni S. Refining at record levels, with National wastewater testing correction factors for P-related arrests nearly Programme Quarter 1 2019. back-calculation of illicit doubling in 5 years. The New Zealand Police, 2019. drug use. Sci Total Envi- New Zealand Herald. 2018. 26. Stats NZ [Internet]. ron. 2016;573:1648-59. Available from: https:// Statistical area 1 dataset 34. Melis M, Castiglioni S, www.nzherald.co.nz/nz/ for 2018 Census – updates Zuczato E. Metabolism and news/article.cfm?c_id=1&- and corrections. 2020. excretion of illicit drugs objectid=12085175 Available from: https:// in humans. In Castiglioni 18. Houlahan M. Meth is www.stats.govt.nz/ S, Zuccato E, Fanelli, R, ‘everywhere’. Otago Daily reports/statistical-area-1- eds. Illicit drugs in the Times. 2017. Available dataset-for-2018-census- environment: Occurrence, from: https://www.odt. updates-and-corrections analysis and fate using co.nz/news/dunedin/ 27. Jack M, Graziadei C. mass spectrometry. New meth-everywhere Report of the Independent Jersey: Wiley, 2011. 19. New Zealand Police Asso- Review of New Zealand’s 35. Exeter DJ, Bronwe M, ciation [Internet]. The war 2018 Census. Welling- Chiang, A, et al. The on meth. Available from: ton: Stats NZ, 2019. 2018 New Zealand https://www.policeassn.org. 28. Lai FY, Ort C, Gartner C, et Index of Multiple nz/news/the-war-on-meth#/ al. Refining the estimation Deprivation (IMD2018): 20. Howell AV, Newcombe DA, of illicit drug consump- Indicators for social and Exeter DJ. The geography of tions from wastewater health research in New methamphetamine manu- analysis: co-analysis of Zealand Brief Report. facture in New Zealand prescription pharmaceu- Auckland: The University between 2004 and 2009. ticals and uncertainty of Auckland, 2020.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 25 www.nzma.org.nz/journal article

36. Boyle C. The rise of 39. UNDOC. World Drug Report Chiaia AC, et al. The spatial cocaine: Drug syndicates 2019. Vienna: UNDOC, 2019. epidemiology of cocaine, ‘determined’ to create a 40. Global Drug Survey [Inter- methamphetamine and market in NZ. The New net]. Global drug survey 3,4‐methylenedioxymeth- Zealand Herald. 2019. 2019. 2019. Available from: amphetamine (MDMA) Available from: https:// https://www.globaldrug- use: a demonstration using www.nzherald.co.nz/nz/ survey.com/gds-2019/ a population measure news/article.cfm?c_id=1&- of community drug load 41. Hutton F. Kiwis, clubs and objectid=12257018 derived from municipal drugs: Club cultures in wastewater. Addiction. 37. Savage J. $20m cocaine Wellington, New Zealand. 2009:104(11):1874-80. case: Eleventh hour Aust N Z J Criminol. admissions of guilt by four 2010;43(1):91-111. 44. Mack AH, Brady KT, men about to stand trial Frances, RJ, Miller, SI, 42. Nefau T, Karolak S, Castillo for smuggling 46kg of drug eds. Clinical textbook L, et al. Presence of illicit into port of Tauranga. The of addictive disorders drugs and metabolites in New Zealand Herald. 2019. (fourth edition). New York: influents and effluents Available from: https:// Guilford Publications, 2016. of 25 sewage water www.nzherald.co.nz/nz/ treatment plants and map 45. Ministry of Health. Drug news/article.cfm?c_id=1&- of drug consumption in use in New Zealand: Key objectid=12244312 France. Sci Total Environ. results of the 2007/08 New 38. EMCDDA. European 2013;461-462:712-22. Zealand alcohol and drug Drug Report. Lisbon: use survey. Wellington: 43. Banta-Green CJ, Field JA, EMCDDA, 2019. Ministry of Health, 2010.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 26 www.nzma.org.nz/journal article

Reducing the MRI outpatient waiting list through a capacity and demand time series improvement programme Heera Bhullar, Bernadette County, Stuart Barnard, Anne Anderson, Mary E Seddon

ABSTRACT INTRODUCTION: A capacity and demand improvement initiative commenced in January 2019 with the goal of reducing the growing outpatient waiting list for magnetic resonance imaging (MRI) at Counties Manukau District Health Board (CMDHB). Initial work showed that the capacity (MRI machines and staff) actually outstripped demand, which challenged pre-existing assumptions. This became the basis for interventions to improve efficiency in the department. Interventions undertaken can be split into three distinct categories: (1) matching capacity to demand, (2) waiting list segmentation and (3) redesigning operational systems. METHODS: A capacity and demand time series during 2019 and 2020 was used as the basis for improving waiting list and operational systems. A combination of the Model for Improvement and Lean principles were used to embed operational improvements. Multiple small tests of change were implemented to various aspects of the MRI waiting list process. Staff engagement was central to the success of the quality improvement (QI) initiatives. The radiological information system (RIS) provided the bulk of the data, and this was supplemented with manual data collection. RESULTS: The number of people waiting for an MRI scan decreased from 1,954 at the start of the project to 413 at its conclusion—an overall reduction of 75%. Moreover, the average waiting time reduced from 96.4 days to 23.1. Achieving the Ministry of Health’s (MoH) Priority 2 (P2) target increased from 23% to 87.5%. CONCLUSION: A partnership between Ko Awatea and the radiology department at CMDHB, examining capacity and demand for MRI and using multiple QI techniques, successfully and sustainably reduced the MRI waiting list over a two-year period. The innovative solutions to match capacity to demand may be instructive for other radiology departments, and other waiting list scenarios.

n 2018 an additional magnetic res- the Ministry of Health (MoH) Priority 2 onance imaging (MRI) machine was (P2) target. The perception was that a lack I purchased, bringing the Counties of medical radiology technicians (MRTs) Manukau District Health Board (CMDHB) was the significant factor preventing the total to three. This increase in physical department from meeting demand. capacity had not reduced the waiting list as In December 2018 the radiology expected and, despite outsourcing 60 scans department requested assistance from Ko per week to private providers, the back- Awatea (CMDHB’s centre for innovation and log and waiting times for MRI scans were improvement) to improve the performance increasing. of the MRI service, particularly to reduce the At the start of the project, 1,954 patients waiting list. were on the waiting list, and only 23% had Stakeholders agreed that a demand MRI scans performed within six weeks— and capacity study would be undertaken

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 27 www.nzma.org.nz/journal article

to identify and realise opportunities to • Average waiting time (measured increase activity. monthly). • Scanning hours utilised per week. Method It was important to measure demand, Staff engagement capacity, backlog and activity in the same All staff groups involved with the MRI units for the same period of time, and to have process participated in its improvement (ie, clear definitions of key metrics (Figure 1). clerical booking staff, nursing, MRTs and radiologists). Regular meetings were held Figure 1: Definitions of terms.3 with staff groups, both separately and collectively, to identify perceived roadblocks; from 1. Demand: What the service is being asked this a framework for improvement was to do developed. Meetings continued regularly: 2. Capacity: What the service could be doing mapping progress, identifying issues and with its resources used optimally defining action plans. 3. Activity: What the service actually did This initiative used Lean tools1 and the 4. Backlog: What the service should have Model for Improvement.2 Lean aims to reduce waste in a system; waste is defined done but haven’t as anything that does not add value (eg, waiting for a test). These tools were used in several ways to: value-stream map the 3. What changes can we make? process, observe how the system actually All improvement requires change and worked and listen and work with front-line being specific about the primary drivers staff. The Model for Improvement uses in managing the MRI waiting list was small tests of change: Plan-Do-Study-Act important. Change interventions (Figure 2) (PDSA) cycles to rapidly trial different ways can be split into three distinct categories: of working. 1. matching capacity to demand The Model for Improvement asks three 2. waiting list segmentation crucial questions that guided the overall 3. redesigning the operational systems. initiative: Interventions were trialled in small tests 1. What are we trying to achieve? of change (PDSA cycles) in each of the three The aim of this study was to optimise areas. the available capacity to better match MRI outpatient demand, reduce the waiting Matching capacity to demand list to less than 500 and meet the MoH’s P2 The first action was to establish the rela- national target of 85% of scans completed tionship between departmental capacity within six weeks. (equipment and staff), incoming demand (referrals) and activity (completed and 2. How will we know that a change is an reported scans). improvement? Demand and activity data were generated Not all change produces improvement— through the radiology information system this question requires the definition of (RIS) reports, which displayed the number measures to confirm improvement: and types of scans being referred and • Number of people on the waiting list completed in chronological order. As the (measured every Tuesday). reports did not record the time each scan • Number of patients waiting in each took, this was added manually. segment (<42 days, 42–90 days, 91–120 Although the MRI machine capacity was days, 121–150 days, 151–180 days, apparent (ie, three MRI machines available >180 days). 24 hours per day), the productive output was • P2 compliance rate—percentage also dependent on the availability of MRTs. of patients scanned within 42 days Capacity data, in the form of staffing rosters from the date of referral (measured by scanner, were translated into available weekly). daily staffing hours.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 28 www.nzma.org.nz/journal article

In reviewing this capacity and demand tions, each machine was staffed with two information, it became clear that there was MRTs. When the new machine arrived and sufficient capacity to meet demand (Figure co-located with the one in the department, 3). In fact, the daily available capacity of this staffing model continued until the staff was greater than incoming demand by staff identified that we could test a ‘2+1 2.5–3 times. model’. This utilised one MRT per scanner Demand was outstripping activity, even and a ‘floating MRT’ shared between two though it was not exceeding the actual rooms, with the focus of ensuring an capacity of the unit. Understanding this efficient flow of patients. This minimised became the basis for interventions to the non-scanning dwell time between increase activity and reduce the waiting list. patients, as the floating MRT could ensure upcoming patients were prepped and ready Demand fluctuated throughout the week, to be scanned as previous diagnostic tests with Mondays and Fridays being the busiest. concluded. The MRT roster was unbalanced and not matched to demand; many part-time MRTs A patient care assistant (PCA) was also were not rostered on Mondays or Fridays, added to the staffing roster—this role was resulting in more MRT capacity than tasked to assist with paperwork, completing scanner availability midweek, and insuf- patient consenting checklists and assisting ficient capacity to run the scanners on the MRTs in getting patients in and out of the busiest days (Figure 4). scanning room. The efficiency of the standalone scanner increased from a nadir of less The 80th percentile of the variation in than 20% to over 99% (Figure 5). the number of hours of incoming demand was chosen to be the minimum number of Scanning list segmentation hours that were required. This meant that a Initial work aimed to decrease the minimum of 18.5 MRT-hours per day (over downtime between scans by grouping scans the three machines) would be required of the same body part (eg, head, shoulder), every weekday to meet demand. Staff were avoiding the need to change MRI coils rostered to be more evenly spread across the between scans and allowing more scans to working week to ensure that each scanner be completed in a list. was fully operational within working hours. Patients with excessive waiting times Historical staffing patterns for each MRI could be grouped into five main groups machine were also reviewed. When there (Figure 6). The average waiting time for this were two MRI machines in different loca- type of patient was close to 300 days.

Figure 2: Driver diagram for MRI optimisation.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 29 www.nzma.org.nz/journal article

Figure 3: Demand and capacity measures over six months.

Figure 4: Capacity and demand in cumulative hours by weekday.

Figure 5: Standalone MRI (Building 58) utilisation.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 30 www.nzma.org.nz/journal article

This knowledge enabled such patients duration scans in addition to the oldest on to be booked onto specific segmented lists. the waiting list. This extracted better value Patients requiring sedation or general from the contractual arrangement. The anaesthetic were scheduled together to outsourcing contract was decreased from 60 enable the anaesthetic workforce to be used scans per week to 15 in July 2019 as part of efficiently. Likewise, patients requiring the DHB cost-saving initiatives. same interpreter language were grouped Redesigning MRI operational and scanned in the same list. The system now has alerts to notify schedulers if certain systems A series of interventions targeted oper- patients are being held up due to one of ational processes, enabling more efficient these characteristics. processes. In March 2019, the service introduced The first intervention was to modify the late weekday sessions, increasing scanning referral vetting process that was creating by two hours per day. At the same time, a bottleneck to workflow. This process was weekend sessions were started. Weekend manual, completed by two senior medical sessions prioritised those patients who had officers (SMOs) and utilised significant been waiting the longest. Pending scans administrative staff time (printing elec- over 180 days were targeted as a priority, tronic referrals for SMOs and scanning with cascading importance being placed on referrals back into the RIS once vetting was subsequent bandings. The waiting list was completed). Referrals that could be vetted also segmented to utilise scanners before by the Grade MRT were identified. The radiologists started work. Unsupervised SMOs’ workload was reduced by redirecting scans were booked at the beginning and end lower-complexity scans while enabling of each day in one-hour blocks, enabling full the Grade MRT to perform at an expanded utilisation of MRT capacity. scope. Furthermore, the Grade MRT used As outsourced scans were performed the electronic system to vet referrals, at a flat-rate fee by private providers, speeding up the process; this encouraged the decision matrix for outsourcing was the SMOs to vet electronically, which in amended to more equally distribute longer

Figure 6: Waiting list segmentation.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 31 www.nzma.org.nz/journal article

turn reduced the workload for adminis- number of cancellations and changes to trative staff. patient bookings. Another operational reform refined the booking template that dictated the duration Results of scanning appointments. The original The waiting list decreased from 1,954 template was provided by the MRI vendor. in January 2019 to 413 in November 2020 However, over time, clinical protocols had (Figure7), and the target compliance for P2 been updated, as had regional and colle- scans increased from 23% to 87.5%, close to giate standards, and these updates were not the MoH’s 90% target. reflected in the booking template. A revised Within the overall backlog reduction, template that accurately reflected up-to-date significant improvements have been made standard scanning times was introduced, in the number of patients waiting in excess allowing the scheduler to accurately book of 42 days (Figure 8). At the commencement based on scan duration times. This resulted of the project, 1,312 patients had waited in efficient booking practices. over 42 days; by the end there were 48, and The allocation of SMOs to MRI sessions the whole waiting list shifted to the left. The was also changed. Initially, when SMOs most dramatic reduction was in the longest were allocated to work in MRI, the admin- wait category, with 204 patients waiting istrative team booked patients according more than 180 days for a scan at the start of to the sub-specialty interest of the SMO (eg, the study, and zero by the end. head and neck patients). This created issues Associated with this change, patients if the SMO was unable to do the session, received MRI scans in a timely manner, with resulting in patients being postponed and 73 days being removed from the average re-booked and wasted scanning capacity. waiting time. Simultaneously, scanning Instead, a patient-focused template was hours per week more than doubled (55.38 to developed enabling sessions to be allocated 136.50)—see Table 1. by patient referral requirement, and SMOs were rostered to cover the sessions. This allowed the roster co-ordinator to allocate Discussion an alternate SMO should the allocated SMO At the start of this study, the radiology be unavailable. This markedly reduced the department had a limited understanding

Figure 7: Number of patients on waiting list, January 2019–November 2020.

Number of Patients on Outpatient MRI Waiting List Jan 2019 - Nov 2020 2000

MRT Strikes MRT Strikes COVID-19 1800 begin. end. Lockdowns begin. 1600

Reduction in 1400 outsourcing numbers

1200

1000

800 Number ofPatients Number 600

400

200

Date

Number on WL Median

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 32 www.nzma.org.nz/journal article

of their capacity and demand for MRI. referrals electronically released SMO time In fact, the department believed more and encouraged SMOs to adopt electronic resources were required, particularly MRTs, vetting). Likewise, understanding the to meet the demand. The collaborative characteristics of the long waits enabled effort between Ko Awatea and the radiology specified lists for those patient groups, department, through this capacity and reducing the long waiting list tail. demand study, showed that in fact there was This attention to detail is not common in sufficient capacity to meet the demand, but waiting list management in New Zealand that it was not organised optimally. healthcare. However, understanding the Through nine interventions covering three principles of Lean thinking (eliminating major areas—matching capacity to demand, waste and increasing value for customers) list segmentation and redesigning opera- has the potential to improve many waiting tional systems—the department sustainably lists. Adopting such manufacturing tools is reduced the number of people waiting for not always appropriate in medicine,4 but MRI scans (from 1,954 to 413), shortened the radiology is perhaps peculiarly suited to this average waiting time (from 96 days to 23 production planning model, as it is a series days) and decreased the number of patients of well-defined technical processes. with excessively long waits (from 1,312 to New Zealand as a whole has relatively few 48). By-products of this ‘shift to the left’ public MRI machines per head of popu- were an improvement in the MoH target lation. In Counties Manukau Health, there for P2 patients (from 23% to >85%) and the are three for a population of 600,000 (~5 exposure of the radiology department to machines per million). It is not clear how QI methodologies and their enthusiasm to many is optimal; internationally numbers continue improvement efforts. range from 55 per million in Japan to The use of data to drive improvement 2.65 in Mexico.5 Given the constrained challenged several long-standing prac- resources, it is important for New Zealand tices (eg, MRTs’ expanded scope to vet to be innovative and apply the appropriate

Figure 8: Changes in waiting list numbers by waiting time segments, December 2018 to November 2020.

Table 1: Waiting list descriptive characteristics.

Descriptive stats December 2018 November 2020 Total number of patients on waiting list 1954 413

Average number of days waiting 96.4 23.1

Range of days waiting 0-308 0–170

Scanning hours per week 55.38 136.50

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 33 www.nzma.org.nz/journal article

improvement methodologies to optimise demand is generalisable. Most radiology resources. departments in New Zealand will be facing A similar approach was used by similar constraints, especially in the face of Canterbury District Health Board (CDHB)6 the COVID-19 lockdowns. when they faced increased waiting times A further constraint is that ethnicity for imaging. Using the principles of Lean, of each patient on the waiting list is not production planning and constraint theories, captured in the RIS, and it is therefore not they worked with in-house production possible to comment on any inequity in planning engineers to improve waiting waiting times. In future it would be bene- times. In this case, the main constraint ficial to conduct this work through an equity was limited radiologist hours, which was lens and, if inequity were to be identified, to improved by rationalising and delegating use patient experience and co-design meth- some tasks traditionally undertaken by odologies to uncover the reasons for this radiologists. As in Counties Manukau Health, disparity. having visibility of the gap between capacity The team faced several constraints that and demand allowed several improvements threatened the sustainability of improve- in the process. ments. The first challenge was the impact A systematic review of the application of of national industrial action that occurred Lean and Six Sigma (which aims to decrease through the third and fourth quarters of defects to one in a million) approaches 2019, severely affecting the availability of in radiology was published in 2016 and MRT staff and directly increasing the waiting concluded that these methodologies had list. At the same time, the outsourcing the potential to reduce errors and cost, contract for 60 scans per week was reduced and improve quality.7 The five studies that to 15 in July 2019, increasing demand. A looked at reducing waiting times were further compounding factor in 2020 was not representative of Counties Manukau COVID-19, which significantly reduced the Health’s situation; starting from a much availability of outpatient scans, particularly shorter baseline: decreasing the waiting during the first lockdown in April 2020— time from 25 days to one. A review of MRI consequently the backlog during this period waiting lists in Canada8 in 2009 noted that increased. most centres routinely used scanners at the weekend, but only 3% were utilised on a Conclusion 24/7 basis, and median scanning hours per Although the original premise for the week was 93.5, whereas our work increased long waiting times for MRI was a lack of this to 136.50. capacity, this study showed that existing There are some limitations to this study. It capacity was sufficient, but inefficiently was neither feasible nor sensible to conduct matched to demand. Acting on detailed a randomised controlled study as the team data of the causes of this inefficiency and were examining the whole department employing Lean thinking principles and and MRI process pathway. Learning and the Model for Improvement methodology, adjusting hypotheses based on small tests of a number of innovative changes were change, a central tenet of most QI methods, made in the process of care, leading to a meant that this was an iterative process with dramatic reduction in waiting times. There multiple tests—some sequential, some in are lessons for other waiting lists in the parallel—so it is unclear which initiative had healthcare system, particularly in building the biggest impact. Therefore, generalising capability in capacity and demand analysis results to other jurisdictions is not possible. and other QI tools, which will be important However, the process of understanding a for the New Zealand health system as it given department’s data on capacity and faces a future of fiscal constriction.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 34 www.nzma.org.nz/journal article

Competing interests: Nil. Acknowledgements: The authoring team would like to acknowledge the co-operation, assistance and involvement from the MRI team at Middlemore Hospital. The staff had a direct positive impact on the uptake and retention of the improvement initiative resulting in improved patient and organisational outcomes. Author information: Heera Bhullar: Improvement Advisor, Ko Awatea, Counties Manukau District Health Board, Auckland. Bernadette County: Portfolio Manager, Ko Awatea, Counties Manukau District Health Board, Auckland. Stuart Barnard: Clinical Director of Central Clinical Services, Counties Manukau District Health Board, Auckland. Anne Anderson: Service Manager, Radiology, Counties Manukau District Health Board, Auckland. Mary E Seddon: Director Ko Awatea, Counties Manukau District Health Board, Auckland. Corresponding author: Dr Mary Seddon, Director Ko Awatea, Middlemore Hospital, Counties Manukau District Health Board, 100 Hospital Road, Auckland 2025, +64275769669 [email protected] URL: www.nzma.org.nz/journal-articles/reducing-the-mri-outpatient-waiting-list-through-a-capacity-and-demand-time-series-improvement-programme

REFERENCES 1. Kruskal JB, Reedy A, Pascal improvement-hub/ and Managing Radiologist L et al. Quality Initiatives. publication/improve- Workload: Application Lean Approach to Improv- ment-leaders-guide-match- of Lean and Constraint ing Performance and ing-capacity-and-de- Theories and Production Efficiency in a Radiology mand-process-and-sys- Planning Principles to Plan- Department. Radiographics tems-thinking/ ning Radiology Services in 2012; 32 (2): 573-587. Accessed 09/02/21 a Major Tertiary Hospital. 2. The Institute for Healthcare 4. Bahensky JA. Lean sigma – Journal of Medical Imaging Improvement [Internet]. will it work for healthcare? and Radiation Oncology. Available from: http:// Journal of Healthcare 2013; 57: 544-550 www.ihi.org/resources/ Information Management 7. Amaratunga A, Dobranows- Pages/HowtoImprove/ 2005; 19 (1): 39 - 44 ki J. Systematic Review of ScienceofImprovemen- 5. Statista [Internet]. 2019. the Application of Lean and tHowtoImprove.aspx Viewed 9th February Six Sigma Quality Improve- Accessed 09/02/21 2021. Available from: ment Methodologies in 3. NHS Institute for Inno- https://www.statista. Radiology. Journal of the vation and Improvement com/statistics/282401/ American College of Radiol- [Internet]. Improvement density-of-magnet- ogy. 2016; 13(9): 1088-1095 Leaders’ Guide – Matching ic-resonance-imag- 8. Emery DJ, Forster AJ, Capacity and Demand ing-units-by-country/ Shojania KG, Magnan S. Process and Systems 6. MacDonald SLS, Cowan IA, Management of MRI Wait Thinking. 2005 https:// Floyd R, et al. Measuring Lists in Canada. Healthcare www.england.nhs.uk/ Policy. 2009; 4 (3):76-86

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 35 www.nzma.org.nz/journal article

Does research help to inform a district health board’s purpose? A qualitative thematic analysis of clinician researcher views Lorraine Neave, Duncan Reid, Brian McKenna

ABSTRACT AIM: The outcomes from research should guide the decisions of healthcare providers, policymakers and funders. This study sought the perspectives of senior hospital clinicians and researchers from a New Zealand district health board (DHB). METHOD: A series of interviews asked participants about the purpose and benefits of research to the DHB, and to reflect upon the enablers and barriers they had experienced in conducting and translating research in a DHB context. RESULTS: Three key themes were identified. The first theme suggested research should inform the DHB’s purpose. The second theme identified how the general busyness, lack of research funding and the differing motivations of clinicians and business leaders doesn’t make it easy to do research in a DHB. The third theme suggested that research barriers could be seen as opportunities. Participants placed importance on an environment that inspires enquiry; that permits staff to stop and question what they do; that overtly informs its community that research is done to improve the delivery of care; that communicates a purposeful research agenda; and that regularly discusses the intersection of research and the purpose of the DHB. CONCLUSION: This study found the absence of an organisation-wide research ethos affected staff engagement in and with research. As a consequence, the effective transfer and translation of knowledge from research was disrupted. Key recommendations were for the DHB to integrate research activity into practice, regularly discuss research evidence and celebrate research achievements.

esearch is fundamental to informing nurture a sense of inter-professional collabo- and improving healthcare outcomes1. ration that is essential to meeting the chang- R A research-enabled environment ing and complex needs of the population.7–9 can engender intellectual curiosity, positive In New Zealand, the Government is the questioning of routine practice and robust major health funder and supports the research practices that can lift process and health of its people through the provision 2,3 practice in general. The outcomes from of a publicly funded, universal healthcare research can also guide healthcare provid- system, with district health boards (DHBs) ers’, policymakers’ and funders’ decisions being the prime recipients of healthcare about resource and purchasing at a national, funding. DHBs are also the prime locality for 4,5 regional and local level. Organisationally, the conduct of the clinical research activity a positive research culture can facilitate in New Zealand.7 Yet, DHBs receive no direct recruitment and retention of excellent clini- capital provision for research in their annual cians and generally improve staff attitudes, government funding, and until 2020 DHBs’ 6 commitment and values. Importantly, it can annual plans to the Minister of Health did not

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 36 www.nzma.org.nz/journal article

require consideration of research. Little is The interviews were thematically known about how DHBs support their staff to analysed utilising a six-stage reflexive manage their research, what things enhance thematic analysis process described by or create barriers to conduct research in Braun and Clarke.10 This comprised listening a DHB and whether the outcomes of that to the recorded interviews multiple times to research are translated to inform practice appreciate the nuanced tones, and reading and support the DHB purpose. the verbatim transcripts for documentary accuracy, before embarking on a recursive Aim process of coding to capture the core semantics. In this process a series of expe- The overarching aim of this study was two riential themes were inductively developed pronged. Firstly, to explore from a DHB-staff and reported in relation to the overarching perspective what the enablers and barriers research question. to doing research were. Secondly, to investigate whether the outcomes of a DHB’s research activity helped to inform that Results DHB’s purpose of providing best care. Eight senior clinicians recognised as research leaders were invited to reflect Method on their experience. Six responded to the invitation, three men and three women, This descriptive, qualitative research was each with decades of experience conducting a part of an explanatory sequential mixed research in a hospital context. The inter- method study that included documentary viewees represented the broad disciplines of analysis, survey and a series of exploratory general medicine, surgery, public health and interviews, reported here. A purposeful psychiatry. Two of the interviewees were sample of clinical leaders who are actively also recognised for their research into indig- researching at a large metropolitan DHB enous health and inequity. All but one had were invited to participate in individual been recipients of, or were key members of, interviews, so we could ascertain their research collaborations that had received experience of conducting research and New Zealand Health Research Council (HRC) translating the outcomes of the research in funding. Two had significant experience the DHB context. Participants consented to conducting industry-sponsored clinical trials the process after having read the relevant in the DHB context. Additionally, all had information sheets outlining the purpose either active or honorary co-appointments of the study. The study was approved by with universities in the region. Auckland University of Technology Ethics Committee (AUTEC), reference 17/204 Three key themes were developed from AUTEC, and the DHB’s locality approval was the data. The first described the participants’ sought and given. universal opinion that research should help to inform the DHB’s purpose. The second The semi-structured interview format theme was struck from a comment that, explored the perceived enablers and barriers. while research should inform the purpose Participants were asked: Do they consider of the DHB, the DHB doesn’t make this that locally conducted research helped to easy. The final theme came from the inter- inform the DHB’s purpose? How had the DHB viewees’ largely optimistic reference to their supported them in their research? Had they being “opportunities” rather than barriers encountered any barriers? And what was for the DHB to support robust research their experience of translating the findings and knowledge creation to inform its of research to practice? The interviews were purpose. Each theme comprised a series of conducted face to face at a pre-arranged sub-themes. time and place suited to the interviewee and were voice recorded. The recordings were Theme 1: research should inform transcribed verbatim, and participants were the DHB’s purpose offered their transcripts to review. Only a The interview participants generally couple of minor clarifications, which did not concurred that research should inform the change the overall context of the interviews, DHB’s purpose (Theme 1), but that the DHB were required. could only do this where ethically and meth-

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 37 www.nzma.org.nz/journal article

odologically sound research was conducted Another commented that more clinicians and was translated to practice. A series of are thinking that “it’s quite cool to work with sub-themes emanated from the interviewee other [non-health] people” who bring their dialogue. diverse and complementary perspectives, In the first sub-theme, interview partic- skills and training. Moreover, research alli- ipants defined the importance of enabling ances with universities means that “we also a culture of questioning within the organ- bring in some young talent through people isation, which “makes a big difference to who are in training” and thereby foster morale” and means “that we don’t just do the DHB’s future workforce. Participants things because we’ve always done it, that did concede that such collaborations aren’t we’re interested in advancing knowledge straightforward and can be complicated by to deliver better care”. They indicated that the need for additional time and effort to while pockets of a research ethos exist in nurture and maintain the relationships. the DHB, that ethos tended to be “enthusiast In the final sub-theme, the interview driven”. Participants felt the DHB could use participants had a collective view that its available resources to research more knowledge creation from research must effectively and more prescriptively. In be disseminated to be translated, and that particular, they said the DHB should use the the intent to disseminate and translate relatively untapped resource of routinely must be planned for from the beginning. collected data focussed on specific health Moreover, translation should be easier issues, and more generally that research where “the research [was conducted] in should inform the DHB’s purpose of the environment it’s going to be trans- improving the broader social determinants lated into” because the end users will have that affect health equity. had the opportunity to “gauge [whether] In the discussion that led to the this [is] going to be something that’s going second sub-theme, interviewees empha- to be useful”. One participant described sised the importance for research to “researching so other people will know”: always be inclusive and real world, “I try and involve the caregivers and the where researchers “work together with community [clinicians] so that they… [are] consumers… having consumers raising the involved and know about what was going [research] questions” because this “leaves on. That makes a big difference”. Partici- people in the organisation confident that pants agreed that, for research to inform the results really are results we should the DHB’s purpose, the DHB needed to listen to”. They considered the outcomes of encourage greater communication about contextually relevant research to be more research plans, create opportunities to share likely to provide decision-makers with the outcomes and overtly celebrate the the confidence that the associated costs translation of research into practices that of translation to practice would meet the inform the DHB’s purpose when it occurs. organisation’s purpose. Theme 2: the DHB environment The third sub-theme reflected on the doesn’t make it easy participants’ views that DHB research In Theme 2 participants provided forth- should be conducted collaboratively and/ right opinions as to why research in a DHB or in partnership with external agencies or is not easy. Four sub-themes explored the bodies. The interview participants univer- barriers encountered. The first reflected sally considered that the breaking down on hospital busyness, which sees research of departmental silos to embrace inter- relegated rather than integrated in the disciplinary and cross-sectoral research organisational ethos. This issue is aggra- collaborations was needed to support the vated by a tendency for staff to resist DHB’s research capacity and capability. changes in their routine, and by business One interviewee noted her own research processes that don’t easily accommodate relationships with universities “bring[s] in timely translation to practice. The overall people with particular expertise around theme was encapsulated in one participant’s design, around models of inquiry, around lament that “the DHB doesn’t make it easy”. analysis”, who are not routinely available in The first sub-theme also described the a DHB. daily grind of hospital busyness being

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 38 www.nzma.org.nz/journal article

a barrier to research. A participant don’t easily accommodate research. Even commented, “it’s very difficult for people to where contextually relevant and well prioritise doing something other than their communicated research has provided the job”, especially where colleagues perceive highest-quality evidence for change, trans- research as a distraction that doesn’t lation can still be stalled by the business directly help to address the immediate process. Participants recognised the restric- issue of “queues of patients at the door”. tions imposed by the government-funding The day-to-day busyness on the frontline model, which frequently constrains the of hospital operations impedes clinicians’ agility of business decision-making. ability to accommodate time for research. However, they were universally critical One participant typified the comments by of the inflexibility they had experienced saying, “you have to do [research] for the when attempting to translate the outcomes love of it more than anything else, and you of research. One account illustrated the have to sort of find the time for it and make frustration: it an important issue for yourself”. Another “We developed a [research] noted, “There’s plenty of interest from programme here with our people, clinicians [in research]… if they had a bit of and we trialled it here with our brain space and a bit of kind of protected people, and the decision-makers time”. were involved in that all the way Translation of new evidence-based through, [but] now trying to get it knowledge to practice should be the corner- implemented we’ve just been going stone of healthcare policy and process. around and around the traps! … However, the second sub-theme described a we’ve got to go through the usual big massive gap between research and clinical business case process, which is quite practice (a ‘know–do gap’).11,12 One partic- time intensive, resource intensive, ipant commented that “in terms of every nobody’s really sure who should be day clinical practice I don’t think there’s writing it, who’s championing it…” an awful lot going on to really try and In the final sub-theme, participants encourage clinicians to be thinking about acknowledged that “not everything is bad putting research into practice”. Others though”. Although conducting research referred to the cause being a general in the DHB was not easy for the various resistance to change, which one partic- reasons outlined, participants noted there ipant described as “partly the conservative are some purposeful pockets of positive nature of our training… and partly it’s just action to support research. Participants simpler to keep doing the way we’ve always noted that the DHB has a research office done because everyone understands their “that you can kind of bounce ideas off role”. Change is seen as “just more work for or [access] some specific skills”. Another people… and you have to spend quite a bit of commented on the research office’s coor- time initially to figure them out before they dination of external grant applications become easier”. “… getting all the right bits together”. One To overcome the know–do gap in research, participant remarked, “the most important participants considered the potential for things have been management’s kind research translation must be nurtured right of acceptance or sometimes even active from the start, when the research question is encouragement of research as something I being developed. This ensures the managers do”. and clinical leaders approving the conduct Theme 3: opportunities, not of the research locally can “gauge [that] this is something that’s going to be useful… barriers In Theme 3 participants focussed on [because] the question itself is framed by the the opportunities moving forward, where service”. The purpose of the research should research can be an integral function of the be well communicated and involve the staff DHB. Two sub-themes detailed how the DHB who may be affected, so that any practice should direct its efforts to enable research change in the future will not be unfamiliar. to better inform its purpose. Firstly, through The third sub-theme addressed a barrier overt communication that research is “what created by DHB business processes that

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 39 www.nzma.org.nz/journal article

we do to improve the care we deliver. And and become a centre of research excellence secondly, by providing competitive seed to support the organisation’s purpose. The funding for research. New Zealand Health Research Strategy 7 The first sub-theme centred on what was 2017–2027 likewise supports such a goal, perceived as a missed opportunity for the stating that a culture of excellence in DHB to overtly inform its community that research is needed to improve the health “research is what we do” to deliver the best and wellbeing of patients and communities. care for everyone. One participant artic- Similarly, the World Health Organization ulated the point best: “I think if we made (WHO) has stated that “universal health [research] more of a deliberate part of what coverage cannot be achieved without we do… it’s not [seen as] our core business evidence from research”, and that to ensure and so we don’t tell patients necessarily this the relevance of research findings to local is what we do… We don’t put up anything context, all nations should both be involved 13 about results anywhere either, even to our as producers and consumers of research. own staff. We’re not very good at saying… Consistent with the literature, participants there’s research going on all the time and in this study considered that research was its purpose is to improve services for our made difficult by the reality of busyness9,14–16 population”. and business process. Differing agendas The second sub-theme discussed the and timelines exist between the hospital importance of a purposeful research budget. business decision makers, who must prior- Interviewees appreciated that the funding itise their decisions to the challenges of their 17–19 the DHB receives from the Government business reality, and the hospital clini- is limited, and that the Government looks cians, who focus on what they know can to the HRC to manage its investment in make a difference for their patients right 3 health research. However, they vociferously now. Participants in this study noted that considered that the government-funding research is frequently left to the enthusiasts model is wrong, observing in other organ- to conduct on top of their business-as-usual isations, both public and private, that a activities. This can lead to clinicians being “percentage of their revenue stream is apathetic towards engagement in and with applied to research and development… research, which won’t be simply overcome But in health… it’s not there”. Interviewees by the implementation of systems to support 20 saw the opportunity for the DHB to target research. Participants considered the research partnerships for bigger projects, inherent mismatch of intents between the and to provide a deliberate competitive organisation’s administrators and clini- research fund for small research projects, cians remains a key barrier to the conduct where seeking external funding would not of research. It was also seen to impact upon be cost or resource effective. Interviewees whether the business accepts the cost– felt if the DHB “can get the ball rolling and benefit of change, and also the effective provide the seed structure and the facilities translation of research outcomes into 14,21 to allow [research] to grow then it will grow practice or policy, where effective dissem- and be self-supporting”. Moreover, such a ination of the findings is dependent upon fund would allow its staff researchers to get the time other health professionals have to the track records needed to apply success- be aware of, agree to, adopt and adhere to a 22 fully for the hotly contested HRC and similar practice or policy change. types of awards. Interview participants in this study concurred that the key to overcome these Discussion barriers is an overt organisational research agenda supported by proactive and early This interview series has provided insight engagement of the users of the ensuing into how a variety of organisational factors knowledge. To this end, they considered influenced staff capacity and capability that the DHB’s executive must deliberately engagement with research. engage with its staff (clinical and business) The executive summary of this DHB’s and potential research collaborators, to research strategy clearly intended for the align their research questions to the DHB’s DHB to build on its research achievements priority healthcare issues. A whole-of-organ-

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 40 www.nzma.org.nz/journal article

isation approach would integrate research conceptual and motivational divide between deeper within the DHB’s core activities and clinical researchers and public health and thereby enable the effective transfer and policy researchers.28,29 translation of research knowledge that is Conclusion integral to the DHB’s purpose. This study has provided insight into A potential limitation of this study was how organisational factors can influence that interviewees were from the medical engagement in and with research and profession, and this may preclude gener- thereby negatively impact the potential for alisation across other health professions. research to inform a DHB’s purpose. These However, the literature suggests nurses factors are all modifiable influences. The similarly describe a lack in the time, themes offer a direction for the DHB’s future training, exposure to and support for research strategy, where research activity research as inhibitors to their engagement is integrated into practice, where research in the research activities.23–27 Likewise, allied evidence is discussed and where research health professionals, and public health achievements are celebrated. The findings and policy evaluation professionals, report from this study will likely resonate across comparable barriers, with the latter noting New Zealand’s DHB landscape. the additional difficulties of navigating the

Competing interests Dr McKenna reports grants from Pacific Health Research Davis Award: HRC, grants from New Zealand Law Foundation, grants from Department of Corrections, grants from Ministry of Health, grants from Marsden Foundation Fund and grants from Borrin Foundation, all outside the submitted work, and that they have a joint professorial position with the Waite- mata District Health Board as a Professor in Forensic Mental Health. Author information: Lorraine Neave: RGON; MHSc; DHSc, Operations Manager, Research and Knowledge Centre, Waitematā District Health Board, Auckland, New Zealand. Duncan Reid: DHSc, MHSc, PgDip HSc, BSc, DipPT, Professor of Physiotherapy, Auckland University of Technology, Auckland, New Zealand Brian McKenna: RN, BA, MHSc, PhD, Professor of Forensic Mental Health, Auckland University of Technology, and the Auckland Regional Forensic Psychiatry Services, Waitematā District Health Board, Auckland, New Zealand; Adjunct Professor, Centre for Forensic Behavioural Science, Swinburne University of Technology, Melbourne, Australia Corresponding author: Dr Lorraine Neave, Research and Knowledge Centre, Waitematā District Health Board, Auckland, New Zealand, Private Bag 93503, +64 9 4868920 ext 42112 [email protected] URL: www.nzma.org.nz/journal-articles/does-research-help-to-inform-a-district-health-boards- purpose-a-qualitative-thematic-analysis-of-clinician-researcher-views

REFERENCES 1. Boaz A, Hanney S, Jones benefits for health-care health.govt.nz/publication/ T, Soper B. Does the performance. Health statement-intent-2015-2019 engagement of clinicians Serv Deliv Res, 2013;1:8. 5. Sarkies MN, Bowles and organisations in 3. Saini V, Garcia-Armesto KA, Skinner EH, et al. research improve health- S, Klemperer D, et al. The effectiveness of care performance: a Drivers of poor medi- research implementation three-stage review? BMJ cal care. The Lancet. strategies for promoting Open 2015;5:e009415. 2017;390(10090):178-90. evidence-informed 2. Hanney, S, Boaz, A, Jones, 4. Ministry of Health [Inter- policy and management T, Soper, B. Engagement net]. Statement of Intent decisions in healthcare: in research: an innovative 2015 to 2019. 2015. Avail- a systematic review. three-stage review of the able from: https://www. Implement Sci, 2017;12:132.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 41 www.nzma.org.nz/journal article

6. Makkar SR, Turner T, whr/2013/report/en/ Implement Sci. 2009;4:46. Williamson A, et al. The 14. Mickan S, Wenke R, Weir 22. Glasziou P, Haynes B. development of ORACLe: K, Bialocerkowski A, The paths from research a measure of an organisa- Noble C. Strategies for to improved health tion’s capacity to engage research engagement of outcomes. Evid Based in evidence-informed clinicians in allied health Nurs. 2005 Apr;8(2):36-8. health policy. Health Res (STRETCH): a mixed 23. Black A, Balneaves Policy Syst. 2016;14:4. methods research protocol. L, Garossino C, et al. 7. Ministry of Business BMJ Open. 2017;7:e014876. Promoting evidence-based Innovation & Employment 15. Pager S, Holden L, Golenko practice through a research and Ministry of Health X. Motivators, enablers, training program for [Internet]. NZ Health and barriers to building point-of-care clinicians. J Research Strategy 2017- allied health research Nurs Adm. 2015;4514-20. 2027. 2017. Available capacity. J Multidiscip 24. Evans G, Duggan R, Boldy from: https://www.health. Health. 2012;5:53-9. D. An exploration of nurs- govt.nz/publication/ 16. Rahman S, Majumder M, ing research perceptions of new-zealand-health-re- Shaban S, et al. Physician registered nurses engaging search-strategy-2017-2027 participation in clinical in research activities at 8. Health Research Council research and trials: issues a metropolitan hospital [Internet]. New Zealand and approaches. Adv Med in Western Australia. Health Research Council - Educ Pract. 2011;7:85-93. Collegian. 2014;21:225-32. About us. 2020. Available 17. Gagliardi A, Berta W, 25. Loke J, Laurenson M, Lee from: https://www.hrc.govt. Kothari A, et al. Integrated K. Embracing a culture nz/what-we-do/about-us knowledge translation in conducting research 9. Royal College of Physicians (IKT) in health care: a requires more than nurses’ [Internet]. Research scoping review. Imple- enthusiasm. Nurse Educ for all. Building a ment Sci. 2016;11:38. Today. 2014;34:132-7. research-active medical 18. Kothari A, MacLean 26. New K, Bogossian F. Nurse workforce. 2016. Avail- L, Edwards N, Hobbs led clinical research: able from: https://www. A. Indicators at the neonatal nurses’ percep- rcplondon.ac.uk/projects/ interface managing tions and experiences. outputs/research-all policymaker researcher Neonatal, Paediatric & 10. Braun V, Clarke V. Using collaboration. Knowledge Child Health Nursing. thematic analysis in Management Research 2012;11:13-18. psychology. Qualitative Practice. 2011;9:203-214. 27. O’Byrne L, Smith S. Models Research in Psycholo- 19. Williamson A, Tait H, to enhance research gy. 2006; 3:77-101. El Jardali F, et al. How capacity and capability 11. Andermann A, Pang T, are evidence generation in clinical nurses: a Newton J, et al. Evidence partnerships between narrative review. J Clin for Health II: Overcom- researchers and poli- Nurs. 2011;20:1365-71. ing barriers to using cy-makers enacted in 28. Huckel Schneider C, Milat evidence in policy and practice? A qualitative A, Moore G. Barriers and practice. Health Res interview study. Health Res facilitators to evaluation Policy Syst. 2016;14:17. Policy Syst. 2019;17:41. of health policies and 12. Lander B, Hanley G, 20. Rahman S, Majumder MA, programs: Policymaker Atkinson-Grosjean J. Shaban SF, et al. Physician and researcher perspec- Clinician-scientists in participation in clinical tives. Eval Program Canada: barriers to career research and trials: issues Plann. 2016;58:208-215. entry and progress. PLoS and approaches. Adv Med 29. Makkar S, Haynes A, One. 2010;5:e13168. Educ Pract. 2011;2:85-93. Williamson A, Redman 13. Dye, C., Boerma, T., Evans, 21. Kothari A, Edwards N, S. Organisational capac- D., et al. Research for Hamel N, Judd M. Is ity and its relationship universal health coverage. research working for you? to research use in six World Health Organization. Validating a tool to examine Australian health policy 2013. Available from: the capacity of health orga- agencies. PLoS One. http://www.who.int/ nizations to use research. 2018;13:e0192528.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 42 www.nzma.org.nz/journal article

How common are non-acute coronary syndrome (ACS) diagnoses in patients with suspected ACS investigated with coronary angiography in New Zealand? (ANZACS-QI 58) Charles Yao-Cheng Ho, Mildred Lee, Seif El-Jack, Peter Barr, Mark Simmonds, Gerry Devlin, Philip D Adamson, Michael Williams, Andrew J Kerr

ABSTRACT BACKGROUND AND AIMS: The last two decades in New Zealand have seen increased availability of primary percutaneous coronary intervention (PCI) for ST-segment elevation myocardial infarction (STEMI) and early invasive coronary angiography (ICA) for other high-risk acute coronary syndrome (ACS) patients. One metric to assess the clinical appropriateness of these invasive strategies is to examine the false-positive rate for the investigation (ie, the rate of non-ACS diagnoses). METHODS: All patients presenting to New Zealand public hospitals with suspected ACS who underwent ICA between 2015 and 2019 were recorded prospectively in the All New Zealand Acute Coronary Syndrome Quality Improvement registry. The cohort was divided according to clinical impression at presentation: (1) suspected STEMI <24h and (2) other suspected ACS. The final discharge diagnosis for each patient were obtained from the registry. RESULTS: There were 6,059 (20%) patients with suspected STEMI <24h and 24,258 (80%) with other suspected ACS. Of the suspected STEMIs <24h, 90.6% had a final diagnosis of STEMI, 3.5% non-ST segment elevation ACS (NSTEACS) and only 5.9% had a non-ACS diagnosis. Of those with other suspected ACS, 80.7% had a final ACS diagnosis. Across all New Zealand district health boards (DHBs), the proportion of non-ACS diagnoses was similar for suspected STEMI presentations. However, for other suspected ACS, the proportions were higher in DHBs with rapid access to coronary interventional facilities than in those without (17.6% vs 7.0%, p<0.001). CONCLUSIONS: False-positive catheter laboratory activations for suspected STEMI patients are low across New Zealand. The differences in the proportion of non-ACS diagnoses according to DHB interventional capability for other suspected ACS requires further investigation.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 43 www.nzma.org.nz/journal article

ver the past two decades, there has investigations, management, in-hospital been a decline in the incidence of outcomes, discharge diagnosis and medi- Oacute coronary syndrome (ACS) and cations. Details of this registry and data associated mortality worldwide, including collection have previously been reported. in New Zealand.1,2 This is in part due to the It is audited monthly to ensure more than increased acute reperfusion therapy for 99% of patients have complete data entry ST-elevation myocardial infarction (STEMI) throughout all New Zealand hospitals, patients presenting within 24 hours of symp- and annual independent audits check the tom onset, using either primary percutane- accuracy of data entry.14,15 ous coronary intervention (PCI) or fibrino- Study cohort lysis with early angiography, and a routine From the ANZACS-QI registry, consecutive invasive approach for those with high risk patients 18 years or older presenting with non-ST elevation ACS (NSTEACS) and other suspected ACS who underwent coronary STEMI patients, both of which have been angiography between January 2015 to shown to improve patient outcomes.2–6 How- January 2019 were included. When patients ever, only 15–30% of patients who present had more than one suspected ACS admis- with chest pain have ACS, and with the sions during the study period, only the first increase in primary PCI for STEMI and early one was included. Patients who were not coronary angiography for NSTEACS, there is New Zealand residents were excluded. The potential for inappropriate invasive coro- study cohort comprised 30,317 patients. nary angiography (ICA) with its associated The cohort was divided into two groups patient and financial risks.7,8 In particular, according to the clinical impression at the out-of-hours activation of catheter labo- the time they entered the catheterisation ratories for suspected STEMIs has a human laboratory: cost for the staff involved in the roster, along with financial costs for the institution in 1. ‘Suspected STEMI <24h’ includes all maintaining the rosters. patients presenting to the catheter laboratory who were suspected of Some studies have reported the rate having a STEMI within 24 hours of of false-positive STEMI diagnosis,9–12 but symptom onset. In New Zealand, no studies have reported specifically on as elsewhere, these patients are false-positive NSTEACS diagnosis and the considered for acute reperfusion appropriateness of ICA in this group of therapy—either a primary PCI or patients, despite NSTEACS comprising fibrinolysis, which is usually followed 60–70% of ACS presentations.3, 13 by early angiography. The choice In this study we aim to identify the inci- depends on how quickly patients dence and characteristics of false-positive can access an interventional cardiac ACS diagnosis and the respective rates of catheter laboratory. In larger metro- non-ACS conditions in patients presenting politan centres most patients are with suspected acute STEMI and other managed with a primary PCI strategy, suspected ACS. but for patients living within district health board (DHB) catchments Method without interventional capability, 16 Data source fibrinolysis is more common. For this analysis the sub-group of those Patients were identified from the All New with suspected STEMI <24h who Zealand Acute Coronary Syndrome Quality underwent a primary PCI strategy was Improvement registry (ANZACS-QI). This is also identified. a nationwide web-based electronic database that captures all patients who present to a 2. ‘Other suspected ACS’ includes all New Zealand public hospital with suspected patients with suspected unstable ACS and who are investigated with coronary angina (UA), non-ST elevation angiography. It records a mandatory dataset myocardial infarction (NSTEMI) and including admission and discharge dates, those with suspected STEMI studied patient demographics, admission ACS risk more than 24 hours after symptom stratification, cardiovascular risk factors, onset.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 44 www.nzma.org.nz/journal article

Definitions mean ± standard deviation. Chi-squared test ‘Final ACS’ or ‘final non-ACS’ are defined or Fisher exact test was used for comparison by the eventual discharge diagnosis, which between two groups, and Wilcoxon Mann– is based on a combination of angiographic Whitney U test or Student’s T-test was used findings, cardiac biomarkers and clinical for continuous variables where appropriate. presentation, at the discretion of the All P-values reported were two-tailed and primary medical practitioner. The diagnosis a P-value <0.05 was considered significant. for each patient and their baseline charac- Data were analysed using SAS statistical teristics were obtained from the ANZACS-QI package, version 9.4 (SAS Institute, Cary, registry. A ‘false-positive’ diagnosis refers NC). to those who were suspected of having ACS initially but were discharged with a non-ACS Results diagnosis. Between 2015 and 2019, a total of 30,317 ‘Myocardial infarction’ (MI) was defined patients 18 years of age or older who according to the contemporary universal presented with suspected ACS were iden- definition.17 UA is diagnosed if one of tified. This comprised 6,059 patients (20%) the following occurred in the absence with suspected STEMI <24h and 24,258 of biochemical evidence of myocardial patients (80%) with other suspected ACS. necrosis: (1) >20 minutes angina pain at rest, Of the 6,059 patients with suspected STEMI (2) de novo Canadian Cardiovascular Society <24h, 5,491 (90.6%) had a final diagnosis class II or III angina or (3) recent destabili- of STEMI, 212 (3.5%) had NSTEACS (191 sation of stable angina with at least CCS class NSTEMI and 21 UA) and 356 (5.9%) had 18 III angina. The final discharge diagnosis non-ACS final diagnoses; 4,072 (67.2%) were for each patient were reported as STEMI, managed with a primary PCI strategy, of NSTEACS (UA and NSTEMI) or non-ACS whom 3,743 (91.9%) had a final diagnosis condition. of STEMI, 61 (1.5%) had a final diagnosis of For the purposes of this analysis, DHBs NSTEACS and 268 (6.6%) had a non-ACS final were dichotomised into those with rapid diagnosis. access to a coronary interventional labo- Of those with other suspected ACS, 19,597 ratory (Waitematā, Auckland, Counties (80.8%) had a final ACS diagnosis, of which Manukau, Bay of Plenty, Waikato, Capital 17,993 (74.2%) were NSTEACS (14,590 and Coast, Hutt Valley, Nelson Marlborough, NSTEMI and 3,403 UA) and 1,604 (6.6%) Canterbury and Southern DHBs) versus were STEMI. There were 4,661 (19.2%) with those without. a non-ACS final diagnosis. Non-invasive testing recorded prior to ICA Regional variation in ACS diagnosis included stress testing (exercise treadmill Among patients suspected of having a test, exercise stress echocardiogram, dobu- STEMI, the rate of false-positive diagnoses tamine stress echocardiogram or nuclear across New Zealand DHBs was less than stress study) and CT coronary angiogram 12%. It appears to be independent of the (CTCA). total number of patients, as demonstrated ‘Obstructive coronary disease on ICA’ in Figure 1. For the DHBs frequently using refers to the presence of coronary stenosis a primary PCI strategy, the false-positive 50% or more in any of the major epicardial rate was low albeit with some variation coronary arteries. (Waitematā 9.9%, Auckland 9.1%, Counties Ethnicity was prioritised using a modified Manukau 8.3%, Bay of Plenty 3.6%, version of New Zealand Standard Ethnicity Waikato 6.7%, Capital and Coast 2.6%, Hutt Data Protocols in the following order: Māori, Valley 2.9%, Nelson Marlborough 3.9%, Pacific, Indian, Other Asian and European/ Canterbury 9.7% and Southern 7.8%, see Other.19 Appendix Table 1). Statistical analysis In comparison, there is a wider range in Descriptive statistics for categorical data non-ACS final diagnoses for other suspected are presented as frequencies and column ACS patients, up to 27.3%. Figure 2 demon- percentages, and continuous variables as strates a general trend for more non-ACS median with interquartile ranges and/or final diagnoses in DHBs with ready access to

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 45 www.nzma.org.nz/journal article

interventional catheter laboratories (17.6% Patients with NSTEACS tend to have more (intervention capable DHBs) vs 7.0% (non-in- cardiovascular comorbidities, including tervention capable DHBs), p<0.001). prior history of MI (12.9% vs 24.1% vs 11.5%, Characteristics and investigations p<0.001), congestive heart failure (CHF) (1.5% vs 4.7% vs 1.1%, p<0.013) and diabetes in patients presenting with mellitus (DM) (17.4% vs 26.4% vs 11.9%, suspected ACS p<0.02). Conversely, those with non-ACS There were 1,200 patients with non-ACS conditions had the lowest burden of cardio- final diagnoses (78 suspected STEMI <24h vascular comorbidities and risk factors. and 1,122 other suspected ACS) who only Patients with a final diagnosis of STEMI had demographic data available in the were intermediate among the other two registry. They were excluded from the subse- groups. Pacific patients had the highest rate quent cohort characteristics comparison, of non-ACS diagnoses compared to other which left 29,117 patients. The demographic ethnicities (Māori 4.6%, Pacific 8.3%, Indian characteristics of this reduced cohort were 5.0%, Other Asian 4.0%, European 4.4%). virtually identical to the original cohort. The frequency of a stress testing or CT Suspected STEMI <24h: In patients coronary angiography (CTCA) performed presenting with suspected STEMI, those prior to undergoing ICA in patients with non-ACS final diagnoses were likely presenting with suspected STEMI was appro- to be younger (median age of STEMI 63 vs priately low (<5%). NSTEACS 63 vs non-ACS patients 59, p<0.007) Other suspected ACS: Patients who and female (26% vs 29% vs 41%, p<0.03). presented with other suspected ACS and

Figure 1: Stacked bar graph demonstrates proportion of final diagnoses in patients presenting to the catheter laboratory with suspected STEMI <24h by DHB (percentages, left-hand y-axis). Scatter plot demonstrates total number of suspected STEMI <24h patients in each DHB (absolute number, right-hand y-axis).

* Rapid intervention capable catheter laboratory available. STEMI, ST-segment elevation myocardial infarction; NSTEACS, non-ST segment elevation acute coronary syndrome; ACS, acute coronary syndrome.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 46 www.nzma.org.nz/journal article

with non-ACS final diagnoses were also obstructive coronary disease (non-ACS younger (median age of ACS 67 vs non-ACS 81.7% vs STEMI 3.5% vs NSTEACS 13.7%, 62, p<0.001) and more likely to be female p<0.001). Findings of significant single- or (31.6% vs 49.3%, p<0.001). Patients with a double-vessel obstructive coronary disease final ACS diagnosis had a higher burden of were more common in those with a final established cardiovascular comorbidities diagnosis of STEMI (75.1% vs 57.6% vs and risk factors (ACS CVD 22.6% vs non-ACS 13.3%, p<0.001). Patients with a final diag- 12.6%, p<0.001). In contrast, those with nosis of NSTEACS had higher proportion of non-ACS conditions were more likely to have three-vessel coronary disease or left-main underlying chronic obstructive pulmonary stenosis than other groups (21.4% vs 28.8% disease (COPD) (9.7% vs 11.5%, p=0.001). vs 5.0%, p<0.001). The proportion with non-ACS diagnoses was A similar trend was demonstrated in similar across all ethnic groups. patients with other suspected ACS. Among The rate of stress testing was 8% (ACS the non-ACS group, 79.6% had no significant 6.5% vs non-ACS 15.9%, p<0.001). CTCA was obstructive coronary disease. performed in 5.5% (ACS 5.4% vs non-ACS Non-ACS diagnoses 6.0%, p=0.206). The non-ACS conditions most likely to Angiographic findings raise suspicion for ACS on presentation Of the patients with suspected STEMI and undergo ICA included stress cardiomy- <24h, the majority of those with non-ACS opathy (11.9%), arrhythmia (10.4%), stable final diagnoses had no significant angina (9.3%) and myocarditis (7.9%). In

Figure 2: Stacked bar graph demonstrates proportion of final diagnoses in patients presenting to the catheter laboratory with other suspected ACS by DHB (percentages, left-hand y axis). Scatter plot demonstrates total number of other suspected ACS patients in each DHB (absolute number, right-hand y-axis).

* Rapid intervention capable catheter laboratory available STEMI, ST-segment elevation myocardial infarction; NSTEACS, non-ST segment elevation acute coronary syndrome; ACS, acute coronary syndrome.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 47 www.nzma.org.nz/journal article

Table 1: Demographics, cohort characteristics and investigations (excludes 1,200 patients without clinical details other than basic demographic information).

Suspected STEMI <24h (n=5981) Other suspected ACS (n=23136) STEMI NSTEACS Non-ACS p-value, p-value, p-value, ACS Non-ACS Final diagnosis n=5491 n=212 n=278 STEMI vs STEMI vs NSTEACS vs n=19597 p-value n=3539 (15.2%) (91.8%) (3.5%) (4.6%) NSTEACS non-ACS non-ACS (84.7%)

Age (years)

<65 n (%) 3,125 (56.9) 123 (58.0) 181 (65.1) 9,145 (46.7) 2,117 (59.8) ≥65 n (%) 2,366 (43.1) 89 (42.0) 97 (34.9) 0.749 0.007 0.109 10,452 (53.3) 1,422 (40.2) <.001 Median (IQR) 63 (54 to 72) 63 (54 to 71.5) 58.5 (47 to 70) 67 (58 to 74) 62 (53 to 71)

Sex, n (%)

Male 4,064 (74.0) 151 (71.2) 164 (59.0) 13,401 (68.4) 1,796 (50.7) 0.365 <.001 0.005 <.001 Female 1427 (26.0) 61 (28.8) 114 (41.0) 6,196 (31.6) 1,743 (49.3)

Ethnicity, n (%)

Māori 542 (9.9) 20 (9.4) 27 (9.7) 2,174 (11.1) 452 (12.8) Pacific 255 (4.6) 21 (9.9) 25 (9.0) 1,012 (5.2) 210 (5.9) Indian 312 (5.7) 14 (6.6) 17 (6.1) 0.010 0.025 0.956 884 (4.5) 136 (3.8) 0.001 Other Asian 247 (4.5) 11 (5.2) 11 (4.0) 592 (3.0) 127 (3.6) European/Other 4,135 (75.3) 146 (68.9) 198 (71.2) 14,935 (76.2) 2,614 (73.9)

Past medical history and risk factors History of CVD, n (%) 1,042 (19.0) 67 (31.6) 38 (13.7) <.001 0.027 <.001 6,874 (35.1) 866 (24.5) <.001

History of MI, n (%) 711 (12.9) 51 (24.1) 32 (11.5) <.001 0.485 <.001 4,432 (22.6) 445 (12.6) <.001

History CHF, n (%) 80 (1.5) 10 (4.7) 3 (1.1) 0.002 0.798 0.013 782 (4.0) 150 (4.2) 0.490

Diabetes, n (%) 955 (17.4) 56 (26.4) 33 (11.9) 0.001 0.017 <.001 4,828 (24.6) 602 (17.0) <.001

COPD, n (%) 395 (7.2) 16 (7.5) 20 (7.2) 0.845 1.000 0.882 1,906 (9.7) 406 (11.5) 0.001

BMI

Mean ± SD 28.5 ± 5.6 29.2 ± 5.8 27.7 ± 6.9 0.111 0.012 0.04 29.4 ± 6.0 29.6 ± 6.8 0.418

Total cholesterol (mmol/l)

Median (IQR) 4.8 (4.0 to 5.7) 4.9 (3.8 to 5.9) 3.9 (3.5 to 4.8) 0.934 0.003 0.004 4.7 (3.8 to 5.7) 4.6 (3.6 to 5.4) 0.079

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 48 www.nzma.org.nz/journal article

Suspected STEMI <24h (n=5981) Other suspected ACS (n=23136) LDL (mmol/l)

Median (IQR) 2.9 (2.2 to 3.6) 2.8 (2.0 to 3.6) 2.4 (1.8 to 3.1) 0.335 <.001 <.001 2.7 (2.0 to 3.6) 2.7 (2.0 to 3.5) 0.211

Smoking status n (%)

Non-smoker 2,364 (43.1) 98 (46.2) 145 (52.2) 8,475 (43.2) 1,891 (53.4) Ex-smoker 1,426 (26.0) 60 (28.3) 59 (21.2) 0.233 0.011 0.183 6,914 (35.3) 1,042 (29.4) <.001 Current smoker 1,701 (31.0) 54 (25.5) 74 (26.6) 4,208 (21.5) 6,06 (17.1)

Clinical presentation Killip Class

I 4,943 (90.0) 192 (90.6) 260 (93.5) 17,838 (91.0) 3,213 (90.8) <.001 0.055 0.225 0.652 II–IV 548 (10.0) 20 (9.4) 18 (6.5) 1,759 (9.0) 326 (9.2)

Admission systolic BP (mmHg)

Median (IQR) 133 135 130 0.565 0.352 0.274 141 138 <.001 (116 to 151) (129 to 150) (114 to 150) (126 to 160) (122 to 155)

Investigations Stress test done, n (%) 46 (0.9) 2 (0.9) 5 (1.8) 0.699 0.097 0.704 1,276 (6.5) 561 (15.9) <.001

CTCA done, n (%) 224 (4.1) 11 (5.2) 8 (2.9) 0.425 0.320 0.189 1,065 (5.4) 211 (6.0) 0.206

LVEF*

Normal (≥50%) 2,087 (38.0) 92 (43.4) 9,543 (48.7) Mild (40 to 49%) 1,409 (25.7) 35 (16.5) - 2,496 (12.7) <.001 - - - - Moderate or severe (<40%) 1,139 (20.7) 32 (15.1) 2,182 (11.1) No EF documented 856 (15.6) 53 (25) 5,376 (27.4)

Revascularisation* 5,045 (91.9) 146 (67.9) - <.001 - - 12,935 (66.0) - -

CABG 239 (4.4) 30 (14.2) - <.001 - - 2,788 (14.2) - -

PCI 4,806 (87.5) 116 (54.7) - <.001 - - 10,147 (51.8) - -

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 49 www.nzma.org.nz/journal article

Suspected STEMI <24h (n=5981) Other suspected ACS (n=23136) Acute reperfusion strategy*

Primary PCI 3,880 (70.7) Thrombolysis 1,174 (21.4) ------None 437 (8.0)

Angiographic findings CAD >50% stenosis on angiogram, n (%)

No obstructive disease 192 (3.5) 29 (13.7) 227 (81.7) 3,025 (15.5) 2,811 (79.6) Single/double vessel disease 4,123 (75.1) 122 (57.6) 37 (13.3) <.001 <.001 <.001 10,438 (53.4) 519 (14.7) <.001 Three vessel disease and/or 1,172 (21.4) 61 (28.8) 14 (5.0) 6,087 (31.1) 203 (5.8) LMS >50%

Time from admission to angiogram (days) n 5487 212 278 19,550 3,533 <.001 0.058 <.001 0.090 Mean ± SD 0.24 ± 0.86 0.75 ± 1.35 0.19 ± 0.78 2.66 ± 2.36 2.66 ± 3.05

STEMI, ST-segment elevation myocardial infarct; NSTEACS, non ST elevation acute coronary syndrome; CVD, cardiovascular disease; MI, myocardial infarction; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; BMI, body mass index; LDL, low density lipoprotein cholesterol; BP, blood pressure ;CTCA, CT coronary angiogram; LVEF, left ventricular ejection fraction; CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention; CAD, coronary artery disease; LMS, left main stem. *Data regarding LVEF and revascularization are not collected for non-ACS patients in the ANZACS-QI registry.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 50 www.nzma.org.nz/journal article

patients suspected of STEMI <24h, stress suggesting that current guidelines are cardiomyopathy (23%), myocarditis (13.3%) generally working well for selecting the and pericarditis (14.4%) were most common, appropriate patients for acute coronary whereas stress cardiomyopathy (11.1%), angiography. However, even in the high- arrhythmia (10.7%) and stable angina (10%) volume centers, there was some variability, were more common in those suspected of with false-positive rates approaching 10% having other ACS (see Appendix Table 2 for in the metropolitan Auckland and Waikato a detailed breakdown). DHBs compared to 3% in the Wellington region. We are unable to determine from Discussion this study whether this DHB variation is because the criteria and processes to This nationwide ANZACS-QI study activate the catheter laboratories are too demonstrates one in twenty patients under- restrictive, resulting in some patients who going ICA for a suspected acute STEMI should be studied early being missed, or <24h had a non-ACS diagnosis. Of the too liberal, resulting in unnecessary acti- sub-group managed with a primary PCI vations with the associated patient risks strategy, the proportion was similar. Of and staff impacts. Our false-positive rate is patients with other suspected ACS, one in relatively low, however, compared to those five had a non-ACS final diagnosis. Half of reported in other series, which can range non-ACS final presentations were for other from 5% to 40%.9-12,20,21 The wide variability cardiac-related conditions. For patients in reported rates can be partially explained investigated with other suspected ACS, the by the different criteria used to define a rate of non-ACS diagnosis was greater in false-positive diagnosis of STEMI. In some DHBs with interventional capability, but for studies, it is based on the absence of angi- those with suspected STEMI the rate was ographically identifiable lesions and/or similarly low across all DHBs. absence of myocardial necrosis biomarkers. Suspected STEMI <24h In others, combinations of angiographic and It is reassuring that the rate of non-ACS clinical data or a discharge diagnosis other diagnosis across the DHBs was low, than STEMI have been used.9,10,12 We defined

Figure 3: Proportion of non-ACS diagnoses according to suspected ACS sub-types.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 51 www.nzma.org.nz/journal article

‘false-positive’ as a discharge diagnosis of of non-ACS diagnoses in DHBs without rapid non-ACS condition. Of the ‘true positives’, access to coronary intervention requires the majority were STEMI, with a very small further investigation, but possible reasons proportion (less than 5%) of other NSTEACS include variation in referral practice and diagnoses. Although these NSTEACS cases greater use of non-invasive investigations also have a high rate of revascularisation, for case selection. Although the use of they would not necessarily require this non-invasive investigation was low in this performed on as urgently. study, the registry does not capture patients The National Cardiac Network and New who had a non-invasive test and were not Zealand ambulance services have recently referred for angiography. developed the National Out-of-Hospital As for suspected STEMI patients, around STEMI Pathway to improve STEMI care.22 half of the non-ACS diagnoses were cardiac This pathway stipulates time frames for related—stress cardiomyopathy (11.1%), reperfusion and recommends criteria arrhythmia (10.7%) and stable angina (10%) for entering the primary PCI pathway. were the common non-ACS diagnoses. Ongoing audit of false-positive cases will A third of patients with a final NSTEACS be important to ensure that patients are diagnosis who had angiography did not not harmed by inappropriate referrals for require revascularisation. There are recent urgent angiography. studies assessing the utility of non-invasive The non-ACS conditions most likely imaging in reducing unnecessary invasive to mimic a STEMI on presentation, and tests. The use of CTCA demonstrated a high therefore to be considered during the acute diagnostic accuracy to rule out significant clinical assessment, included stress cardio- coronary artery disease in patients with myopathy (23%), myocarditis (13.3%) and NSTEACS in the VERDICT trial.24 Likewise, pericarditis (14.4%). ICA is appropriately initial cardiac MRI or CTCA appears to performed as a diagnostic tool for these reduce unnecessary invasive coronary conditions that are associated with marked angiography in CARMENTA.25 However, the ECG changes and symptoms that can mimic RAPID-CTCA trial was unable to show any potentially life-threatening STEMI. benefit in early CTCA utilisation when hard Other suspected ACS end-points such as death and myocardial infarction were assessed.26 Non-invasive In contrast, there is a paucity of prior data imaging as a routine diagnostic tool in on false-positive rates in other suspected suspected ACS is likely to evolve in the near ACS presentations. We have previously future. reported the overall rate of coronary angiography for ACS in New Zealand to be Demographic considerations lower in comparison to similar European We found that women suspected of having countries.1–3 This is even more pronounced a STEMI or other ACS were more likely to in non-interventional DHBs.23 Although have a non-ACS diagnosis,9,10,12 which is at 19.2% the non-ACS rate for suspected similar to the findings reported in other other ACS in our study is higher than for studies. This may reflect the known greater suspected STEMI, it is still much lower difficulty in making ACS diagnoses due to than the proportion of patients presenting atypical and equivocal symptoms more with chest pain who have a final diagnosis often observed in women.5,18 Moreover, of ACS. This is consistent with the use of stress cardiomyopathy, the most common diagnostic pathways to select those more non-ACS diagnosis in our study (suspected likely to have ACS for invasive angiography. STEMI 23%, other suspected ACS 11.1%), These patients have less well defined ECG occurs predominantly in women, which may changes and more cardiovascular comor- contribute to the gender differences.5,17,18 bidities, in particular pre-existing history of Non-ACS diagnoses were also more common MI and diabetes mellitus, which may lower in younger patients. This may represent the threshold for ICA in equivocal presen- their lower prevalence of atherosclerotic tations. Because of its role as a non-urgent disease relative to other conditions. It is also diagnostic test, a higher false-positive rate is possible that clinicians are more likely to appropriate for this group of patients. The refer these patients when the presentation is reasons for the observed lower proportion equivocal.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 52 www.nzma.org.nz/journal article

Ethnic differences in false-positive STEMI referred for ICA. All MIs have been included diagnoses in non-Europeans have been as final ACS diagnoses for this study. There reported on internationally, such as higher is a small proportion of NSTEMI cases (6%) rates of false-positive STEMI in African recorded as Type II MI in the cohort. Whilst Americans in the Activate-SF Registry.10 In reallocation of Type II MIs to the non-ACS our study, Pacific patients appear to have group would increase the false-positive rate, a higher rate of false-positive diagnosis this may be misleading because the cases in comparison to other ethnicities when clinically selected for invasive coronary presenting with a suspected STEMI, but angiography often requires it to make the not in suspected other ACS. A recent study diagnosis and a third of these patients by Grey et al demonstrated that ischaemic received revascularisation for stable heart disease (IHD) mortality for Māori and coronary artery disease. Although the final Pacific people remains twice that for Euro- ACS and broad non-ACS diagnoses were peans in New Zealand, despite an overall available and reported for all patients, the declining trend in IHD deaths and hospi- more detailed non-ACS diagnoses and other talisations across all ethnic groups—albeit data were not captured. This occurred in a this decline was smaller among Pacific small number of DHBs, making systematic people.27 It is therefore conceivable that bias unlikely. the diagnostic threshold for interpreting an equivocal presentation or ECG as a STEMI in Conclusions Pacific patients may be lower. However, it is Patients who present with suspected ACS unclear why this was not observed in Māori and have a final non-ACS diagnosis were patients and suspected other ACS patients in more likely to be younger, female and our study. This highlights a perspective for have fewer cardiovascular comorbidities. future studies to assess diagnostic differ- False-positive catheter laboratory activa- ences between ethnic groups in ACS. tions for suspected STEMI <24h patients Limitations are generally low across New Zealand. The The study cohort was patients with difference in the proportion of non-ACS suspected ACS who were referred and diagnosis in other suspected ACS presenta- underwent ICA. We have no information tions between DHBs with and without rapid regarding how many patients with ACS access to coronary interventional facilities might have been missed due to the condition requires further investigation. not being suspected and/or not being

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 53 www.nzma.org.nz/journal article

Competing interests: Nil. Author information: Charles Yao-Cheng Ho: Department of Cardiology, Counties Manukau District Health Board, New Zealand. Mildred Lee: Department of Cardiology, Counties Manukau District Health Board, New Zealand. Seif El-Jack: Department of Cardiology, Waitematā District Health Board, New Zealand. Peter Barr: Department of Cardiology, Auckland District Health Board, New Zealand. Mark Simmonds: Department of Cardiology, Capital and Coast District Health Board, New Zealand. Gerry Devlin: Department of Cardiology, Gisborne Hospital, New Zealand. Philip D Adamson: Department of Cardiology, Canterbury District Health Board, New Zealand; Department of Medicine, University of Otago, New Zealand. Michael Williams: Department of Cardiology, Southern District Health Board, New Zealand; Department of Medicine, University of Otago, New Zealand. Andrew J Kerr: Department of Cardiology, Counties Manukau District Health Board, New Zealand; Department of Medicine, University of Auckland, Auckland, New Zealand. Corresponding author: Dr Charles Yao-Cheng Ho, c/o Department of Cardiology, Middlemore Hospital, Otahuhu, Auckland 93311, New Zealand [email protected] URL: www.nzma.org.nz/journal-articles/how-common-are-non-acute-coronary-syndrome-acs-diagnoses-in-patients-with-suspected-acs-investigated-with-coronary-angiography-in-new-zealand-anzacs-qi-58

REFERENCES 1. Wang TKM, Grey C, Heart Journal. 2019; 40. 8. Lindsell CJ, Anantha- Jiang Y, et al. Trends in 5. Ibanez B, James S, Agewall raman V Fau - Diercks cardiovascular outcomes S, et al. 2017 ESC Guidelines D, Diercks D Fau - Han after acute coronary for the management of JH, et al. The Internet syndrome in New Zealand acute myocardial infarction Tracking Registry of Acute 2006–2016. Heart. 2020. in patients presenting Coronary Syndromes 2. Timmis A, Townsend N, with ST-segment elevation: (i*trACS): a multicenter Gale CP, et al. European The Task Force for the registry of patients with Society of Cardiology: management of acute suspicion of acute coro- Cardiovascular Disease myocardial infarction in nary syndromes reported Statistics 2019. Euro- patients presenting with using the standardized pean Heart Journal. ST-segment elevation of the reporting guidelines for 2019; 41:12-85. European Society of Cardi- emergency department chest pain studies. Annals 3. Benjamin EJ, Muntner ology (ESC). European Heart of Emergency Medicine. P, Alonso A, et al. Heart Journal. 2017; 39:119-77. Disease and Stroke 6. ST-elevation myocardial 9. Regueiro A, Fernandez-Ro- Statistics-2019 Update: A infarction: New Zealand driguez D, Freixa X, et Report From the American Management Guidelines, al. False Positive STEMI Heart Association. Circula- 2013. New Zealand Medical Activations in a Regional tion. 2019; 139:e56-e528. Journal. 2013; 126:127-64. Network: Comprehensive Analysis and Clinical 4. Wang TKM, Grey C, Jiang 7. Launbjerg J, Fruergaard Impact. Results From the Y, Jackson R, Kerr A. P, Hesse B, Jorgensen Catalonian Codi Infart Epidemiology of acute F, Elsborg L, Petri A. Network. Rev Esp Cardiol coronary syndrome by Long-term risk of death, (Engl Ed). 2018; 71:243-9. subtype in New Zealand cardiac events and 2006–2016: an ANZACS-QI recurrent chest pain in 10. McCabe JM, Armstrong EJ, nationwide linkage study of patients with acute chest Kulkarni A, et al. Preva- hospitalisation, procedures pain of different origin. lence and factors associated and case fatality. European Cardiology. 1996; 87:60-6. with false-positive ST-segment elevation myocardial

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 54 www.nzma.org.nz/journal article

infarction diagnoses at datasets (ANZACS-QI 25). Out-of-Hospital STEMI Path- primary percutaneous New Zealand Medical way. New Zealand National coronary intervention-ca- Journal. 2019; 132:19-29. Cardiac Network, 2018. pable centers: a report 16. Kerr A, Lee M, Grey C, et 23. Williams MJA, Harding SA, from the Activate-SF al. Acute reperfusion for Devlin G, et al. National registry. Arch Intern ST-elevation myocardial variation in coronary Med. 2012; 172:864-71. infarction in New Zealand angiography rates and 11. Lu J, Bagai A, Buller C, et (2015-2017): patient and timing after an acute al. Incidence and charac- system delay (ANZACS-QI coronary syndrome in teristics of inappropriate 29). The New Zealand medi- New Zealand (ANZACS-QI and false-positive cardiac cal journal. 2019; 132:41-59. 6). New Zealand Medical catheterization laboratory 17. Thygesen K, Alpert JS, Journal. 2016; 129:66-78,6. activations in a regional Jaffe AS, et al. Fourth 24. Linde JJ, Kelbæk H, Hansen primary percutaneous universal definition of TF, et al. Coronary CT coronary intervention myocardial infarction Angiography in Patients program. Am Heart J. (2018). European Heart With Non-ST-Segment 2016; 173:126-33. Journal. 2018; 40:237-69. Elevation Acute Coro- 12. Larson DM, Menssen 18. Roffi M, Patrono C, nary Syndrome. J Am Coll KM, Sharkey SW, et al. Collet J-P, et al. 2015 Cardiol. 2020; 75:453-63. “False-Positive” Cardiac ESC Guidelines for the 25. Smulders MW, Kietselaer B, Catheterization Laboratory management of acute Wildberger JE, et al. Initial Activation Among Patients coronary syndromes in Imaging-Guided Strategy With Suspected patients presenting without Versus Routine Care in ST-Segment Elevation persistent ST-segment Patients With Non-ST-Seg- Myocardial Infarction. elevation: Task Force for ment Elevation Myocardial JAMA. 2007; 298:2754-60. the Management of Acute Infarction. J Am Coll 13. Townsend N, Wilson L, Coronary Syndromes in Cardiol. 2019; 74:2466-77. Bhatnagar P, Wickra- Patients Presenting without 26. Gray AJ, Roobottom masinghe K, Rayner M, Persistent ST-Segment C, Smith JE, et al. The Nichols M. Cardiovascular Elevation of the European RAPID-CTCA trial (Rapid disease in Europe: epidemi- Society of Cardiology (ESC). Assessment of Potential ological update 2016. Eur European Heart Journal. Ischaemic Heart Disease Heart J. 2016; 37:3232-45. 2016; 37:267-315. with CTCA) - a multicentre 14. Kerr A, Williams MJ, 19. Health Mo. HISO parallel-group randomised White H, et al. The All New 10001:2017 Ethnicity Data trial to compare early Zealand Acute Coronary Protocols. Wellington: computerised tomography Syndrome Quality Ministry of Health., 2017. coronary angiography Improvement Programme: 20. Lange DC, Rokos IC, Garvey versus standard care in Implementation, Meth- JL, Larson DM, Henry TD. patients presenting with odology and Cohorts False Activations for ST-Seg- suspected or confirmed (ANZACS-QI 9). New ment Elevation Myocardial acute coronary syndrome: Zealand Medical Journal. Infarction. Interv Cardiol study protocol for a 15. Kerr AJ, Lee M, Jiang Y, et Clin. 2016; 5:451-69. randomised controlled trial. Trials. 2016; 17:579. al. High level of capture of 21. Lange DC, Conte S, Pappas- coronary intervention and Block E, et al. Cancellation 27. Grey C, Jackson R, Wells S, associated acute coronary of the Cardiac Catheteriza- et al. Trends in ischaemic syndromes in the all New tion Lab After Activation heart disease: patterns Zealand acute coronary for ST-Segment-Elevation of hospitalisation and syndrome quality improve- Myocardial Infarction. Circ mortality rates differ by ment cardiac registry and Cardiovasc Qual Outcomes. ethnicity (ANZACS-QI excellent agreement with 2018; 11:e004464. 21). New Zealand Medical national administrative Journal. 2018; 131:21-31. 22. New Zealand National

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 55 www.nzma.org.nz/journal article

Antipsychotic and sedative medication use in long-term care facilities providing dementia care Etuini Ma’u, Janine Burton, Elizabeth Fussell

ABSTRACT AIM: This study aimed to quantify use of antipsychotic and sedative medications in residents with dementia in long-term care facilities in the Waikato District Health Board (DHB) catchment and to identify factors associated with the prescription of these medications. METHODS: Resident records and the medication charts of 271 residents with a diagnosis of dementia from 13 dementia units in the Waikato DHB catchment, as well as the psychogeriatric (PG) unit, were reviewed for current prescriptions for any antipsychotic or sedative medication and the date those medications were most recently prescribed. RESULTS: Antipsychotics were prescribed for 133 (49.1%) residents, with a mean (95% CI) of 401 (354–448) days since the most recent prescription was made. Only 16.8% of antipsychotic prescriptions were prescribed in the preceding 12 weeks, with 31.3% of prescriptions prescribed more than a year prior. Residents were more likely to be prescribed an antipsychotic if they were male (56.9% vs 42.6%, p=.019) or had an incident form completed (30.8% vs 19.6% p=.03). Regression analysis showed only gender (OR 1.79, 95%CI 1.07–2.98, p=.026) was associated with antipsychotic medication prescription. Sedatives were prescribed for 60 (22.1%) residents, with a mean (95%CI) of 487 (431–544) days since the most recent prescription was made, and 44.8% of prescriptions were dated more than a year prior. Residents were more likely to be prescribed a sedative if they entered the facility at a younger age (76.9 vs 79.5, p=.042) or had been in the current facility longer (980 vs 734 days, p=.048). Following regression analysis, no individual factors were significantly associated with sedative prescription. CONCLUSION: With clear evidence of the risks of antipsychotics to patients with dementia, the proportion of residents prescribed an antipsychotic or sedative in this study, in conjunction with the prolonged duration of prescription, is cause for concern and needs addressing.

ementia, a syndrome characterised behaviour frequently occurring in people by progressive decline in cognition with dementia. BPSD is common, with an and daily functioning, increases estimated point prevalence of 60–80% and D 2 in prevalence with advancing age. As the a cumulative risk of 95%. The resultant proportion of the New Zealand population distress of BPSD causes both individual suf- aged 65 years and older increases, there fering and increased caregiver burden.3 It will be a concomitant increase in dementia is also associated with worsening cognition cases. Cases of dementia are projected to and a hastened progression to a more severe triple to 170,000 by the year 2050.1 This in dementia.4 turn will increase the prevalence of those The causes of BPSD are usually multifac- presenting with the unique challenges torial in nature, with a complex interplay associated with the behavioural and psy- of both biological and psychosocial factors chological symptoms of dementia (BPSD). contributing to both aetiology and patho- BPSD is defined by the International Psy- genesis. As such, BPSD treatment guidelines5 chogeriatric Association (IPA) as symptoms emphasise a holistic understanding of the of disturbed thoughts, mood, perception or patient, their symptoms and the circum-

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 56 www.nzma.org.nz/journal article

stances within which the symptoms occur. level care require a diagnosis of dementia to The mainstay of management centres be eligible, it has been estimated that 55% to around psychosocial interventions, with 77% of those in the other types of care also the use of psychotropic medications only have a diagnosis of dementia.15 considered when trials of behavioural or Most ARC facilities are privately owned psychological interventions have not been enterprises contracted by district health of adequate benefit and the risks associated boards (DHBs) to provide an agreed number with medication use have been weighed and type of residential care beds.16 Facilities 5 up. Evidence for the efficacy of antipsy- operate to a national standard of services, chotic medication in BPSD is poor, with 5–11 but variability remains, including in the use patients needing to be treated to achieve of psychotropic medication, as a result of a clinically significant improvement in one combination of individual facility character- 6 additional patient. Furthermore, recent istics and organisational culture.17 Studies studies have demonstrated that imple- investigating the prescription of psycho- menting an antipsychotic deprescribing tropic medications in a New Zealand setting protocol after 12 weeks on an antipsychotic are heterogenous, with some focussed on is not associated with any increase in BPSD prescribing rates in the general population 7,8 symptoms in 67–80% of patients. and others on psychotropic prescribing in The risks associated with the use of aged residential care. antipsychotics are significant and include Three studies investigate psychotropic cerebrovascular events, gait disturbance, medication prescribing in the general popu- orthostatic hypotension, sedation, QT lation aged 65 years and older. Ndukwe 9 prolongation and increased mortality. et al18 conducted a population-level study There is also evidence of an association using pharamaceutical dispensing data to between duration of antipsychotic use and describe and characterise the prescription 10 mortality, with an additional death in 100 of psychotropic medication in the 65+ patients with BPSD treated for 12 weeks population between 2005 and 2013. They increasing to 17 additional deaths in 100 demonstrated a 27.5% increase in the 11 patients with BPSD treated for two years. dispensing of a defined daily dose per These risks resulted in an FDA black box 1,000 older people per day (DDD/TOPD) warning in 2005 against the use of anti- for antipsychotics, a 10.3% increase in the 12 psychotics in patients with dementia. DDD/TOPD for sedatives/hypnotics, and a Despite these known risks and the black 5.9% decrease in the DDD/TOPD for anxio- box warning, use of antipsychotics and lytics. In the same period there was a 22.5% benzodiazepines in residential care remains increase in the prescription of any psycho- common. Indeed, a recent systematic tropic medication, with the proportion of 13 review and meta-analysis reviewing 89 psychotropic prescriptions accounted for international studies of central nervous by antipsychotics increasing from 4.3% to system medication use in aged residential 4.5% and decreasing for both hypnotics/ care calculated an overall pooled estimate sedatives (37.2% to 33.5%) and anxiolytics of 26.1% for antipsychotic and 36.2% for (7.1% to 5.5%). Jackson et al10 used the New benzodiazepine use. Zealand Atlas of Healthcare variation to When no longer able to be safely cared for examine the dispensing of antipsychotics at home, long-term care for those aged 65 and benzodiazepines in the New Zealand years and older in New Zealand is provided 65+ population between 2008 and 2012. by one of four types of aged residential care Across the four-year period of the study, (ARC) facilities. Rest homes and hospital antipsychotic prescription increased from level care provide care for those requiring 22.4 to 23.8 per 1,000 people, and benzodiaz- physical assistance with activities of daily epine prescription increased from 106 to 109 living. Dementia care is utilised by those per 1,000 people. The study also commented who need a secure facility but have limited on a relationship between antipsychotic physical care needs. Psychogeriatric (PG) dispensing and dementia/psychogeriatric-oc- care, the highest level of care, is reserved cupied bed days but did not present data for those with complex mental, cognitive or to support this statement. Using Ministry physical needs.14 Although those in dementia of Health prescribing data, Wilkinson and

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 57 www.nzma.org.nz/journal article

Mulder19 also showed a 49% increase in in 2013 following a range of interventions antipsychotic prescribing across all ages to improve prescribing practices. Peri et between 2008 and 2015. They demonstrated al17 investigated 537 residents across 14 that rates of prescribing by ethnicity were residential care facilities (including four the highest across both genders in the 65+ dementia units) and found antipsychotics age group for Europeans (male 3.97% and prescribed in 20%, hypnotics in 31% and female 5.04%), Pacific peoples (male 3.51% sedatives in 19% of all residents. and female 3.61%) and Asian (male 1.68% These studies indicate the prescription of and female 2.53%) and for Māori females antipsychotics in New Zealand has at best (4.09%). Rates of antipsychotic prescribing remained unchanged, if not increased, since in Māori males aged 65+ (3.50%) were 1990. The prevalence of antipsychotic use is second only to Māori males in the 25–44- further magnified in dementia and psycho- year-old age group (4.77%). geriatric care facilities where the majority Five New Zealand studies investigate of residents have a diagnosis of dementia. psychotropic medication prescribing in The aim of this study was to quantify use aged residential care. Kerse20 carried out a of antipsychotic and sedatives in dementia medication chart and medical record audit units and the PG unit in the Waikato District of 606 residents in 14 randomly selected Health Board (WDHB) catchment area and aged residential care facilities in 1999–2000, potentially identify factors associated with 50% of whom had a diagnosis of dementia. the prescription of psychotropic medication. Benzodiazepines were prescribed in 33.5% of all residents and ‘major tranquilisers’ Methods in 17% of residents. Compared to those This study was approved by the New without dementia, residents with a diag- Zealand Ministry of Health’s Health & nosis of dementia were more likely to Disability Ethics Committee, approval be prescribed a major tranquiliser (25% number 16/STH/135. vs 10%, p<.001). Tucker and Hosford21 surveyed the prescription of psychotropic Setting and participants medications in 1,053 residents from 26 rest Dementia units and the PG unit in the homes (including five dementia units and WDHB catchment area were invited to 11 private hospitals) in 2005 and compared participate in the study in December 2016. the findings to a similar survey carried out These facilities are all privately owned in 1990. They showed an increase in anti- enterprises, contracted to WDHB to provide psychotic prescribing from 21.9% to 23.7%, aged residential care beds at dementia or PG with a higher level of antipsychotic use in level care. PG is the highest level of resi- dementia units (59.5%) compared with 17% dential care and is required for dementia for rest homes and 29.5% for hospital level residents with more complex physical and/ care. Benzodiazepine prescribing over the or behavioural care needs that cannot be same period reduced from 29.6% to 12.4%. adequately provided in a dementia unit. Heppenstall et al 22 reviewed the electronic Spanning both rural and urban areas, the records of a stratified random sample of WDHB is one of 20 district health boards 222 residents from the 2008 Older Persons’ in New Zealand, with a 65+ population of Ability Level (OPAL) study of all aged resi- 65,000. At the time of the study there were dential care facilities within the Auckland 18 dementia units and a psychogeriatric region’s three district health boards. They hospital in the region. All 14 dementia units reported antipsychotics prescription in within Hamilton City and the Waikato, 19.1% of residents, benzodiazepines in South Waikato, Matamata-Piako, and 21.0% and a non-benzodiazepine hypnotic Waipa districts were approached. Due (zopiclone) in 12.5%. Tordoff et al 23 to geographical practicalities, the four undertook an audit of 228 residents across dementia units situated further afield in the 13 dementia and psychogeriatric units in Thames/Coromandel, Hauraki and Ruapehu 2011 and found antipsychotic prescription districts were not approached. Thirteen in 50.4% and benzodiazepine prescription of the 14 dementia units approached, as in 39%. Prescribing had reduced to 38.2% well as the PG unit, agreed to participate. and 25.8% respectively when remeasured Resident records and the medication charts

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 58 www.nzma.org.nz/journal article

of all 280 residents in participating units average duration of 791 (706–877) days. were reviewed. Following exclusion of nine Sixty-eight residents (25.1%) had at least one residents with primary diagnoses of schizo- incident form completed in the preceding phrenia, bipolar disorder or an intellectual six months. disability, 271 residents with a primary Antipsychotics were prescribed for diagnosis of dementia were included in the 133 residents for a mean (95% CI) of 401 analysis. (354–448) days, giving a point prevalence Data collection of 49.1% (43.0%–55.2%). Sedatives were Baseline demographic data were collected prescribed for 60 residents for a mean of for all residents and included age, gender, 487 (431–544) days, giving a point preva- ethnicity, diagnosis, age at entry and lence of 22.1% (17.3%–27.6%). Among those duration of stay in their current facility. prescribed an antipsychotic, 16.8% of the Each resident’s current medication chart current prescriptions were issued in the was reviewed for current prescription preceding 12 weeks, with 31.3% of prescrip- of any antipsychotic and/or sedative tions unadjusted for >52 weeks. Among medications. Antipsychotic medication those prescribed a sedative, 19% were included both first- and second-gener- issued their prescriptions in the preceding ation antipsychotics. Sedative medications 12 weeks, with 44.8% of prescriptions unad- included all benzodiazepines and zopi- justed for >52 weeks. clone, a non-benzodiazepine hypnotic. The Resident characteristics of those date the medication was most recently prescribed antipsychotics to those who were prescribed was used as a proxy for the not are compared in Table 2. There was duration (in days) since the medication no difference in resident age (80.1 vs 81.9 was last reviewed. All resident files were years, p=.072), age at entry into their current also reviewed for formal incident forms dementia unit (78.2 vs 79.6 years, p=.187), or completed by staff in the preceding six duration of residence in their current unit months, as a proxy for BPSD in excess of (727 vs 849 days, p=.169). Residents were what would reasonably be expected in a more likely to be prescribed an antipsy- secure dementia facility. chotic if they were male (56.9% vs 42.6%, Data analysis p=.019) or had an incident form completed Continuous data are summarised by (30.8% vs 19.6% p=.034). A binomial logistic means and standard deviations (SD) and regression was performed to ascertain ranges and counts by numerators and the effects of age, age at entry into care, denominators and proportions expressed as duration in current care home, gender and percentages. For the continuous variables, having an incident form on the likelihood normality assumptions were reasonably of being prescribed an antipsychotic (Table well met, and t-tests and ANOVA were used 3), with only male gender shown to be a as appropriate for comparisons of groups. predictor of antipsychotic prescription (OR For categorical variables, Chi-square tests of 1.79, 95%CI 1.07–2.95, p=.026). independence were used to compare groups. Resident characteristics comparing those Binomial logistic regression was used to prescribed sedatives to those who were assess confounding between independent not are presented in Table 4. There was variables. Statistical analysis was carried no difference in resident age (79.6 vs 81.4 out using the Statistical Package for Social years, p=.140), gender (24.4% vs 20.3%, Sciences version 20.0 (SPSS Inc., Chicago, Il, p=.416) or having had an incident form USA). completed (31.7% vs 23.2% p=.183). Resi- dents were more likely to be prescribed Results a sedative if they entered the facility at a younger age (76.9 vs 79.5 years, p=.042), Demographic details are presented in had been in the current facility longer (980 Table 1. Residents had a mean (95% CI) age vs 734 days, p=.048) or were resident in of 81 (80.0–81.9) years, 54.6% were female psychogeriatric hospital (36.6% vs 15.9%, and 90.8% were New Zealand European/ p<.001). A binomial logistic regression (Table Pākehā. Mean age of entry into their current 5) showed no factor significantly predicted home was 78.9 (77.9–79.9) years, for an sedative prescription.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 59 www.nzma.org.nz/journal article

Table 1: Demographic characteristics of residents.

Total (N=271)

Age (years) Mean 81.0 95% CI 80.0–81.9

Age at entry into facility (years) Mean 78.9 95% CI 77.9–79.9

Gender (%) Male 123 (45.4) Female 148 (54.6)

Ethnicity (%) New Zealand European/Pākehā 246 (90.8) Māori 10 (3.7) Other 15 (5.5)

Duration at current facility (days) Mean (SD) 791 95% CI 706–877

Residents with Number of residents (%) 68 (25.1) incident forms (%)

Antipsychotic (%) N (%) 133 (49.1)

Antipsychotic Mean 401 duration (days) 95% CI 354–448

Antipsychotic duration (%) <12 weeks 22 (16.8) 12–52 weeks 68 (51.9) >52 weeks 41 (31.3)

Sedative (%) n(%) 60 (22.1)

Sedative duration (days) Mean 487 95% CI 431–544

Sedative duration (%) <12 weeks 11 (19) 12–52 weeks 21 (36.2) >52 weeks 26 (44.8)

Level of care (%) Dementia care 189 (69.7) Psychogeriatric 82 (31.3)

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 60 www.nzma.org.nz/journal article

Table 2: Characteristics of residents prescribed an antipsychotic compared to those who were not.

Prescribed anti- Not prescribed Sig. psychotic antipsychotic N=133 N=138

Age (years) Mean 80.1 81.9 95% CI 78.6–81.6 80.6–83.3 .072

Age at entry into Mean 78.2 79.6 care (years) 95% CI 76.7–79.3 78.2–81.0 .187

Duration in Mean 727 849 current 95% CI 703–878 722–977 .169 home (days)

Gender Female 63 (47.4) 85 (61.6) Male 70 (52.6) 53 (38.4) .019

Ethnicity New Zealand European/Pākehā 121 (94.5)) 125 (93.3) Māori 5 (3.9) 5 (3.7) Other 3 (1.6) 4 (3.0) .945

Incident forms Yes (%) 41 (30.8) 27 (19.6) .034

Prescribed a sedative Yes (%) 40 (30.1) 20 (14.5) .002

Table 3: Univariate (unadjusted OR) and logistic regression (adjusted OR) comparing residents prescribed an antipsychotic to those who were not.

Unadjusted OR Sig. Adjusted OR Sig. (95%CI) (95%CI) Age 0.97 (0.95–1.00) .073 0.96 (0.53–1.73 .892

Age at entry 0.98 (0.95–1.01) .187 1.02 (0.57–1.84) .942

Gender 1.78 (1.10–2.89) .019 1.79 (1.07–2.98) .026

Incident form 1.82 (1.05–3.20) .034 1.72 (0.97–3.22) .062

Duration in 0.99 (0.99–1.00) .172 0.99 (0.99–1.00) .908 current facility

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 61 www.nzma.org.nz/journal article

Table 4: Characteristics of residents prescribed a sedative compared to those who were not.

Prescribed sed- Not prescribed Sig. ative sedative N=60 N=211

Age (years) Mean 79.6 81.4 .140 95% CI 77.3–82.0 80.3–82.5

Age at entry into care (years) Mean 76.9 79.5 95% CI 74.5–79.5 78.4–80.6 .042

Duration in current home (days) Mean 980 734 95% CI 753–1,208 643–825 .020

Gender Female 30 (50) 118 (55.9) .416 Male 30 (50) 93 (44.1)

Ethnicity New Zealand European/Pākehā 53 (89.8) 193 (94.6) Māori 3 (5.1) 7 (3.4) Other 3 (5.1) 4 (2.0) 0.347

Incident forms Yes (%) 19 (31.7) 49 (23.2) 0.183

Prescribed an Yes (%) 40 (66.7) 93 (44.1) .002 antipsychotic

Table 5: Univariate (unadjusted OR) and logistic regression (adjusted OR) of comparing residents prescribed a sedative to those who were not.

Unadjusted OR Sig. Adjusted OR Sig. (95%CI) (95%CI) Age 0.97 (0.94–1.01) .140 .98 (0.49–1.99) .972

Age at entry 0.97 (0.93–1.00) .044 .99 (0.49–2.00) .980

Gender 1.27 (0.71–2.25) .417 1.29 (0.70–2.38) .411

Incident form 1.54 (0.82–2.88) .185 1.61 (0.85–3.11) .148

Duration in 1.00 (1.00–1.00) .023 1.00 (0.99–1.00) .649 current facility

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 62 www.nzma.org.nz/journal article

and regression analysis showed only male Discussion gender (OR 1.78x, p=.026), not a completed Despite the increasing evidence around incident form (OR 1.72, p=.062), was asso- the relatively poor efficacy and increased ciated with antipsychotic prescription. Dual risks associated with antipsychotic and prescribing of antipsychotics and sedatives sedative prescription in individuals with was common, with 40 (14.8%) residents dementia, this study shows the use of these prescribed both. Residents prescribed an medications in dementia care and PG facil- antipsychotic more than twice as likely to be ities remains high. Although there was a prescribed a sedative compared to those not trend towards reduced prescribing of anti- prescribed an antipsychotic (30% vs 14.5%, psychotics in dementia units, from 59.5% in p=.002). 200521 to 50.4% in Auckland in 2011,23 the While it is reassuring other demographic 48.2% of residents in this study currently factors have not been shown to be influ- prescribed an antipsychotic suggests this encing antipsychotic prescribing, it is trend has plateaued. concerning that medication prescription The mean duration since the prescription was not associated with the completion of of either an antipsychotic or sedative was incident forms, which are usually completed last adjusted was also high, at 401 and 487 when residents manifest BPSD in excess days for antipsychotics and sedatives respec- of what would usually be expected for tively. Only 16.8% of those prescribed an their level of care. Only 30.8% of those antipsychotic and 19% of those prescribed prescribed an antipsychotic and 31.7% of a sedative had the most recent dose change those prescribed a sedative had at least one occur within the recommended 12-week incident form for behavioural concerns period.5 The use of the date of prescription completed in the preceding six months. It is as a proxy for duration since the medication possible that the behaviour of some resi- dose was last reviewed is acknowledged as dents may have been noted to be escalating a limitation of this study, as it is possible and was therefore dealt with before the there was a review of the rationale for the threshold for an incident form was reached. medication in the preceding weeks and It is also possible the presence of symptoms an active decision made not to make any severe enough to warrant an incident form adjustment. It is also a possibility that the being completed may have resulted in an last medication change reflected a dose active decision not to make any psychotropic decrease and that the resident was being medication changes for some residents. actively monitored before consideration of However, over two-thirds of residents a further dose reduction. However, 31.3% of prescribed an antipsychotic or sedative those on antipsychotics and 48.4% of those had not manifested behaviours meeting the on sedatives had not had their prescriptions threshold for completing an incident form in adjusted for over a year, far exceeding the the preceding six months and could poten- 12-week recommendation and suggesting tially have trialled a medication decrease. that reviews are not actively occurring. A strength of this study is that 13 of the Residents prescribed antipsychotics were 14 dementia units approached, and the PG more likely to be to male and have had unit, agreed to participate in the research, an incident form completed in the last six reducing the risk of selection bias and months. Those prescribed sedatives were providing an accurate cross-sectional more likely to be younger and to have snapshot of antipsychotic and sedative been in their current home for longer. The medication use as of December 2017 in the finding that antipsychotic prescriptions Waikato DHB catchment area. As described were more likely to be for male residents above, there are limitations to both using the (1.79x) and those who had an incident form date of medication prescription as a proxy completed (1.82x) likely reflects the fact that for duration since the medication dose was the behaviour of these residents was inter- last reviewed and using incident forms as preted as more threatening, and that the a proxy for BPSD severity. However, the assessed risk of harm from that behaviour is proportion of residents who had not had higher. However, gender was also associated incident forms completed or had their with having an incident form completed, medications altered for over 12 months is

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 63 www.nzma.org.nz/journal article

significant and suggests that active medi- needs addressing. Recent evidence that cation reviews are not occurring. antipsychotic deprescribing regimens7,8 are With clear evidence of the risks of anti- not associated with an escalation in BPSD psychotics to patients with dementia,6 indicate that active and regular reviews the proportion of residents prescribed an of prescribing, in conjunction with the 23 antipsychotic or sedative in this study, in upskilling of care staff, can significantly conjunction with the prolonged duration reduce unnecessary antipsychotic and of prescription, is cause for concern and sedative prescription in this population.

Competing interests: Nil. Acknowledgements: This research was funded by a summer studentship research grant from the Waikato Clinical School, University of Auckland. The authors would like to thank the participating aged residential care facilities for their involvement. Author information: Etuini Ma’u: Senior Lecturer, University of Auckland. Janine Burton: Clinical Nurse Specialist, Waikato District Health Board. Elizabeth Fussell: Consultant Psychiatrist, Waikato District Health Board. Corresponding author: Etuini Ma’u, Department of Psychological Medicine, Peter Rothwell Academic Centre, Waikato Clinical Campus, University of Auckland, Private Bag 3200, Hamilton 3240 Etuini.ma’[email protected] URL: www.nzma.org.nz/journal-articles/antipsychotic-and-sedative-medication-use-in-long-term-care-facilities-providing-dementia-care

REFERENCES 1. Deloitte Access Economics. decline in Alzheimer’s Assoc. 2018;19(11):1009-14. Dementia Economic Impact disease: data from the 8. Brodaty H, Aerts L, Harri- Report 2016. Welling- ICTUS study. J Neurol. son F, Jessop T, Cations ton: Alzheimers New 2013;260(7):1859-65. M, Chenoweth L, et al. Zealand; 2017. 5. RANZCP [Internet]. Anti- Antipsychotic Deprescrip- 2. Steinberg M, Shao H, Zandi psychotic medications as a tion for Older Adults in P, Lyketsos C, Welsh- treatment of behavioural Long-term Care: The HALT Bohmer K, Norton M, et and psychological symp- Study. J Am Med Dir Assoc. al. Point and 5-year period toms of dementia 2016. 2018;19(7):592-600 e7. prevalence of neuropsy- Available from: https:// 9. Steinberg M, Lyketsos CG. chiatric symptoms in www.ranzcp.org/files/ Atypical antipsychotic dementia: the Cache County resources/college_state- use in patients with Study. Int J Geriatr Psychi- ments/practice_guidelines/ dementia: managing safety atry. 2008;23(2):170-7. pg10-pdf.aspx concerns. Am J Psychiatry. 3. Feast A, Moniz-Cook E, 6. Banerjee S. The use of 2012;169(9):900-6. Stoner C, Charlesworth antipsychotic medication 10. Jackson G, Gerard C, Minko G, Orrell M. A systematic for people with dementia: N, Parsotam N. Variation review of the relationship Time for action. 2009. in benzodiazepine and between behavioral and 7. Van Leeuwen E, Petrovic antipsychotic use in people psychological symptoms M, van Driel M, De Sutter aged 65 years and over (BPSD) and caregiver A, Stichele R, Declercq T, in New Zealand. N Z Med well-being. Int Psychogeri- et al. Discontinuation of J. 2014;127(1396):67-78. atr. 2016;28(11):1761-74. Long-Term Antipsychotic 11. Ballard C, Hanney 4. Canevelli M, Adali N, Drug Use for Behavioral ML, Theodoulou M, Cantet C, Andrieu S, and Psychological Symp- Douglas S, McShane R, Bruno G, Cesari M, et toms in Older Adults Aged Kossakowski K, et al. The al. Impact of behavioral 65 Years and Older With dementia antipsychotic subsyndromes on cognitive Dementia. J Am Med Dir withdrawal trial (DART-

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 64 www.nzma.org.nz/journal article

AD): long-term follow-up M, Ballard C, Bisla J, Calem 19. Wilkinson S, Mulder R. of a randomised place- M, et al. Current prevalence Antipsychotic prescribing bo-controlled trial. Lancet of dementia, depression in New Zealand between Neurol. 2009;8(2):151-7. and behavioural problems 2008 and 2015. N Z Med 12. Food and Drug Adminis- in the older adult care J. 2018;131(1480):61-7. tration [Internet]. Atypical home sector: the South 20. Kerse N. Medication use antipsychotic drugs. Avail- East London Care Home in residential care. New able from: http://www. Survey. Age Ageing. Zealand Family Physi- fda.gov/Safety/MedWatch/ 2014;43(4):562-7. cian. 2005;32(4):251. SafetyInformation/ 16. Technical Advisory Services 21. Tucker M, Hosford I. Use SafetyAlertsforHumanMed- (TAS) [Internet]. Aged of psychotropic medicines icalProducts/ucm150688. residential care (ARC). in residential care facilities htm Available from: https:// for older people in Hawke’s 13. Hasan SS, Zaidi STR, tas.health.nz/dhb-pro- Bay, New Zealand. N Z Med Nirwan JS, Ghori MU, Javid grammes-and-contracts/ J. 2008;121(1274):18-25. health-of-older- F, Ahmadi K, et al. Use of 22. Heppenstall CP, Broad JB, people-programme/ Central Nervous System Boyd M, Hikaka J, Zhang X, aged-residential-care/ (CNS) Medicines in Aged Kennedy J, et al. Medica- Care Homes: A Systematic 17. Peri K, Kerse N, Moyes S, tion use and potentially Review and Meta-Analysis. Scahill S, Chen C, Hong inappropriate medications J Clin Med. 2019;8(9). J, et al. Is psychotropic in those with limited prog- 14. Ministry of Health [Inter- medication use related nosis living in residential net]. Residential care to organisational and aged care. Australas J questions and answers. treatment culture in resi- Ageing. 2016;35(2):E18-24. dential care. J Health Organ Available from: https:// 23. Tordoff J, Ailabouni N, Manag. 2015;29(7):1065-79. www.health.govt.nz/ Browne D, Al-Sallami H, our-work/life-stages/ 18. Ndukwe H, Tordoff Gray A. Improvements health-older-people/ J, Wang T, Nishtala P. in the prescribing of long-term-residential-care/ Psychotropic medicine antipsychotics in dementia residential-care-ques- utilization in older people and psychogeriatric units tions-and-answers in New Zealand from in New Zealand. Int J Clin 15. Stewart R, Hotopf M, Dewey 2005 to 2013. Drugs Aging. Pharm. 2016;38(4):941-9. 2014;31(10):755-68.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 65 www.nzma.org.nz/journal article

Social impacts and costs of schizophrenia: a national cohort study using New Zealand linked administrative data Sheree Gibb, Naomi Brewer, Nicholas Bowden

ABSTRACT AIM: To identify a national population of individuals living with schizophrenia in New Zealand, and to examine health, social support, justice, economic outcomes and estimated government costs compared to a matched comparison group. METHODS: Data were sourced from the Integrated Data Infrastructure. Individuals with a schizophrenia diagnosis in public hospital discharge or specialist secondary mental health service data, aged 18 to 64 and living in New Zealand were included in the schizophrenia population. Propensity score matching was used to select a comparison group of individuals without schizophrenia from the New Zealand resident population and compare outcomes and costs. RESULTS: In 2015 there were 18,096 people living with schizophrenia in New Zealand, a prevalence of 6.7 per 1,000 people. Compared to the matched comparison population, individuals with schizophrenia had higher hospitalisation rates for mental (OR=52.80) and physical (OR=1.18) health conditions. They were more likely to receive social welfare benefits (OR=17.64), less likely to be employed (OR=0.11) and had lower income ($26,226 lower). Per-person government costs were higher for the schizophrenia group across all domains, particularly health ($14,847 higher) and social support ($11,823 higher). CONCLUSION: Schizophrenia is associated with a range of adverse health, social and economic outcomes and considerably higher government costs compared to the general population.

chizophrenia is a chronic and disruptive health conditions.9–11 The early adult onset illness and has been ranked one of the and chronic course of schizophrenia has Stop 25 causes of disability globally.1 The a considerable impact on economic and illness typically presents in the early adult social outcomes including employment, years and has a chronic course, usually re- income, housing and contact with the justice quiring lifelong treatment and monitoring.2–6 system.12,13 For example, in an international Estimates of schizophrenia prevalence review, employment rates among adults with 14 vary worldwide, with a recent meta-analysis schizophrenia ranged from 11.9% to 39.0%, finding a median lifetime prevalence of and a recent study found that the cumulative 0.48% (interquartile range 0.34–0.85%)7 earnings among working age individuals and another finding a median lifetime with schizophrenia was equivalent to only 15 prevalence of 0.40% (interquartile range 14% of those without schizophrenia. 0.30–0.66%).8 Despite the relatively low These adverse outcomes for individuals prevalence, the health, social, economic and living with schizophrenia are associated personal costs associated with schizophrenia with a range of direct and indirect financial are substantial. Individuals with schizo- costs. A recent systematic review of inter- phrenia are at increased risk for premature national studies found annual costs for death, both from suicide and from physical the schizophrenia population ranged

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 66 www.nzma.org.nz/journal article

from US$94 million to US$102 billion comparison group who do not have (2013 dollars), and the economic burden schizophrenia of schizophrenia was estimated to range 3. construct broad estimates of from 0.02% to 1.65% of gross domestic government costs for individuals with 16 product. In the United States, estimated and without schizophrenia across a excess direct healthcare costs attributed to range of sectors. schizophrenia in 2013 were US$37.7 billion while indirect costs totaled US$117 billion.17 Similarly, in England, societal costs of Methods schizophrenia were estimated at £6.7 billion Data source: the Integrated Data in 2004/05, including direct government Infrastructure costs of £2 billion.13 In Australia, costs of Data were sourced from Stats NZ’s IDI. psychosis to society have been estimated The IDI is a large database containing at AU$4.91 billion annually (as of 2010), of linked individual-level microdata about which AU$3.52 billion were borne by the people and households in New Zealand.24 Australian government.18 Data come from a range of government and There is limited New Zealand epidemio- non-government administrative and survey logical research on schizophrenia. Studies data sources including health, tax, welfare using administrative data on health service and justice. All data are probabilistically use have suggested a 12-month incidence linked and de-identified; only approved of 0.5 per 1,000 for Māori and 0.3 per 1,000 researchers can access data in a secure for non-Māori,19 and a 12-month preva- environment; and all outputs must be aggre- lence of 1.0 per 1,000 for Māori and 0.3 per gated, confidentialised and checked by Stats 1,000 for non-Māori.20 Other studies have NZ before release. examined various clinical aspects of schizo- For more information about IDI, see phrenia illness and treatment.21–23 However, https://www.stats.govt.nz/integrated-data/ there has been little research on the social integrated-data-infrastructure. and economic outcomes and associated Participant population government costs for people living with The base population used for this study schizophrenia in New Zealand. was the IDI estimated resident population Previously, the investigation of the life of New Zealand (IDI-ERP) for 31 December outcomes of schizophrenia at a national 2015, created using established methods level has been limited by the lack of for identifying a resident population within available social and economic measures the IDI.25,26 This population aims to capture in administrative health service data. The all individuals who were alive and living recent development of New Zealand’s in New Zealand as at 31 December 2015. Integrated Data Infrastructure (IDI) now All individuals in the schizophrenia and provides an opportunity to quantify the comparison populations were drawn from health, social support, justice sector and this population. economic outcomes for individuals with Individuals were included in the schizo- schizophrenia in New Zealand.24 The IDI also phrenia population if they had a diagnosis of contains a range of cost information that can schizophrenia, defined as: potentially be used to estimate government costs, although methods for using these costs 1. having a public hospital discharge are underdeveloped. in the National Minimum Dataset (NMDS) between the dates of 1 The aims of this paper are to use linked January 1988 (date the dataset begins) data from the IDI to: and 31 December 2015 with a primary 1. identify a national population of indi- or secondary diagnosis code for viduals living with schizophrenia in schizophrenia using International New Zealand in 2015 Classification of Disease (ICD)-9 or 2. examine the health, social support, ICD-10 (see Appendix Figure 1 for list justice sector and economic outcomes of codes), or for individuals with schizophrenia, 2. having a record in the Programme and to contrast those with a for the Integration of Mental Health

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 67 www.nzma.org.nz/journal article

Data (PRIMHD) diagnosis dataset for to identify individuals who received one of secondary mental healthcare between the Tier 1 working age benefits (Jobseeker the dates 1 July 2008 (date the dataset Support, Supported Living Payment and Sole begins) and 31 December 2015 with Parent Support). a diagnosis code for schizophrenia using Diagnostic and Statistical Social housing Manual of Mental Disorders (DSM)-IV, Kāinga Ora data were used to identify ICD-9 or ICD-10 (see Appendix Figure individuals who were an applicant on a 1 for list of codes). social housing tenancy agreement. For comparison, a cohort of individuals Socioeconomic deprivation: the New Zealand without schizophrenia was identified. The Index of Deprivation 2013 absence of a diagnosis code for schizo- Address notification data were used to phrenia in either NMDS or PRIMHD over the identify each individual’s meshblock of resi- same time periods defined this group. This dence and the corresponding New Zealand cohort was matched to the schizophrenia Index of Deprivation 2013 (NZDep2013) population on age, sex and ethnicity (see score. NZDep2013 measures socioeconomic the subsection Analytical sample for further deprivation based on the meshblock (small details). area geography comprising approximately The schizophrenia and comparison popu- 100 households) in which individuals live.28 lations were further restricted to individuals It is available for everyone who has an aged 18 to 64 as at 31 December 2015 to address recorded in the IDI (over 99% of the capture the typical working age population, sample). Deprivation scores were collapsed because many of the outcome measures are into deciles, with 1 representing the least only relevant for working age individuals, deprived and 10 the most deprived. and because mental health care for ages 65 and over is not well recorded in PRIMHD.27 Justice sector Outcome measures Police proceedings Unless otherwise stated all outcomes were New Zealand Police data were used estimated for the 2015 calendar year. to identify individuals who had police proceedings initiated against them Health (including warrants, arrests and charges, Public hospital admissions but excluding infringement notices). NMDS data were used to identify admis- Conviction sions to a public hospital. Ministry of Justice data were used to Inpatient psychiatric admission identify individuals who were convicted of PRIMHD data were used to identify admis- a crime. sions to an inpatient psychiatric facility. Imprisonment Emergency department visits Department of Corrections data were National Non-Admitted Patient Collection used to identify individuals who were (NNPAC) data were used to identify atten- incarcerated. dance at an emergency department. Economic Accident Compensation Corporation support Income Accident Compensation Corporation (ACC) Inland Revenue (IR) data were used to data were used to identify individuals who calculate the total taxable income from: had received ACC support through either wages and salaries; withholding payments; income compensation or having their government benefits (including Jobseeker medical expenses paid for by ACC. Support, Supported Living Payment, Sole Parent Support and New Zealand Superan- Social support nuation); student allowance; paid parental Social welfare benefit received leave; pensions (superannuation); claimants The Ministry of Social Development compensation; and declared self-em- (MSD) Benefit Dynamics Dataset was used ployment income.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 68 www.nzma.org.nz/journal article

Income estimates do not include ACC costs were calculated by summing investment income, cash income or income payments made for medical expenses (from earned overseas, as these sources of income ACC records) and compensation for income are not reliably recorded in IR records. loss (from IR records). All medical expenses costs were allocated to the year in which the Employment accident occurred. Compensation was allo- IR data were used to identify employed cated to the year in which the compensation individuals, defined as all those who was paid. received income from wages, salaries or self-employment. Social support Benefit costs were calculated by summing Costs the Tier 1 (as previously defined) benefit All costs were calculated for the 2015 payments from the IR tables (which record calendar year in New Zealand dollars. actual amount paid) and Tier 2 (payments Unless otherwise stated, costs were drawn and allowances such as housing supple- from the Social Investment Analytical Layer ments and child disability allowances) (SIAL) (https://swa.govt.nz/what-we-do/ entitlements from MSD Benefit Dynamics analytics/social-investment-analyti- Dataset tables (which record entitlement: cal-tool/) produced by the Social Investment actual amount paid is not available for Tier Agency, which calculates individual-level 2 benefits). government service costs within the IDI. Social housing costs included total Health government rent subsidies for social Health costs included government housing, calculated from the Kāinga Ora costs for: hospital admissions; emergency tables in the IDI. The total social housing department (ED) attendances; outpatient cost was allocated to the person who was visits; community pharmaceuticals; labo- the main applicant on the social housing ratory tests; general medical subsidies; application. Therefore, individuals living specialty mental health services; and ACC. in social housing where they were not the They did not include primary healthcare, main applicant on the application (eg, if they private healthcare or patient co-payments. were the partner of the main applicant) will Hospital admission costs were calculated not have social housing costs recorded. by multiplying the case mix cost weight by the purchase unit cost for the relevant year, Justice sector Justice costs were taken directly from with no further adjustment. Outpatient and the Ministry of Justice and Department of ED costs were calculated from purchase Corrections data collections in the IDI. Costs unit costs in the NNPAC dataset. The general included government costs for charges, medical subsidy and the government court proceedings, imprisonment and other subsidies for laboratory tests and sentences. Where a sentence lasted more community pharmaceuticals were obtained than a year, only the current year of costs directly from the community laboratory or was included. pharmaceutical table. Mental health costs were calculated using Total government costs purchase unit costs obtained from the Social This was calculated as the sum of all Investment Agency, which were applied health, social support and justice sector to events in the PRIMHD dataset. These costs. included all publicly funded secondary Demographics mental health consultations and treatments, Age (18–24, 25–34, 35–44, 45–54, 55–64 including both inpatient and outpatient years), sex (male/female) and ethnicity were general, forensic and addiction services. It sourced from the personal detail table in the included pharmaceuticals given in inpatient IDI. Ethnicity was recorded in total response settings but excluded private treatment, format (so an individual can identify with community dispensed pharmaceuticals more than one ethnic group) and, for (which are included under the pharma- this study, restricted to Level 1 groupings ceuticals cost category above) and mental (European, Māori, Pacific, Asian). The health treatment given in primary care. Middle Eastern, Latin American and African

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 69 www.nzma.org.nz/journal article

(MELAA) and Other ethnic categorisations individuals without schizophrenia. Nearest had small counts and were therefore not neighbour one-to-one matching was under- used in this study. taken using the MatchIt package version Procedure 3.0.2 in R. Matches were drawn without replacement from the IDI-ERP for 2015. Data were accessed from the September Matching variables were age at the end of 2017 refresh of the IDI. Data were extracted 2015, sex and ethnicity. These variables using SAS version 7.129 and analysed using were selected because they were available R version 3.6.130 and Stata/MP version 15.31 for almost all individuals and rarely change Random rounding to base 3 was applied to over time. Individuals who had ever been all counts, in adherence to Stats NZ confi- diagnosised with schizophrenia or who dentiality requirements. had missing data on any of the matching Statistical analysis variables were not eligible to be selected as Calculation of prevalence matches. Schizophrenia prevalence was calcu- We also conducted two sensitivity analyses lated as the proportion of the New Zealand with variations to this matching method in IDI-ERP of 18 to 64 year olds for 2015 who which each individual with schizophrenia had a schizophrenia diagnosis identified was matched to five and ten controls any time until 31 December 2015. Preva- respectively. Results were not substantively lence was calculated overall and by sex, different to the single matching, so we age and ethnicity. Odds ratios (ORs) were elected to stay with single matching for the generated using logistic regressions for analysis (see Appendix Tables 1–6). each sociodemographic sub-group. ORs for ethnicity were adjusted for age and sex to Comparisons of outcomes and costs take account of differences in population To compare outcomes between the schizo- structure between ethnic groups. phrenia and comparison groups we fitted linear (for continuous outcomes) or logistic Analytical sample (for dichotomous outcomes) regression The analytical sample for the outcomes models in which the outcome was modelled analysis contained a schizophrenia group as a function of schizophrenia population defined as all those in the New Zealand membership. Estimated odds ratios (for IDI-ERP of 18 to 64 year olds for 2015 with dichotomous outcomes) and beta coefficients a schizophrenia diagnosis for at least two (for continuous outcomes) were drawn from years on or prior to 31 December 2013, and these models. who had complete data on the matching Ethics approval variables of age, sex and ethnicity. We This study was approved by the University restricted the sample to individuals with at of Otago Human Research Ethics Committee least a two-year history of schizophrenia in (Reference: H17/062). The study was order to exclude impacts that are related to reviewed as a ‘Minimal Risk Health Research the period of initial diagnosis. To examine – Audit and Audit related studies’ proposal. whether a longer delay from diagnosis to Access to IDI data was approved by Stats NZ. outcomes might be required, a sensitivity analysis was run on an alternative schizophrenia population whose diagnosis was Results recorded on or prior to 31 December 2010 Schizophrenia population size and (indicating they had been living with schizo- prevalence phrenia for at least five years). Results were A total of 18,096 individuals aged 18 to 64 not substantively different to those from the had a schizophrenia diagnosis recorded on main analysis (see Appendix Tables 1–6), so or prior to 31 December 2015. This is equiv- we opted to use the two-year restriction as it alent to a prevalence of 6.7 per 1,000 people, retains a larger sample size. or 0.67%. For comparison, a matched sample Table 1 shows the number and corre- without schizophrenia was also generated sponding prevalence of individuals with from the IDI-ERP of 18 to 64 year olds for schizophrenia in New Zealand by age, sex 2015. Propensity score matching methods and ethnicity. Schizophrenia prevalence were used to select the comparison group of

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 70 www.nzma.org.nz/journal article

was significantly lower among females phrenia and non-schizophrenia sub-groups compared to males (OR 0.52) and higher in being identical on the demographic Māori and Pacific ethnic groups compared measures, so for the purposes of presenting to non-Māori and non-Pacific (ORs 3.36 and the analytical sample, the groups have 1.70 respectively). Prevalence was highest in been combined. The majority (63.8%) of the the middle adult years (35 to 54) and lowest analytical sample were male. Approximately in the youngest (18 to 24). two thirds (65.4%) identified as European, Description of the population for one third (35.4%) as Māori, 11.2% as Pacific and 5.6% as Asian. The majority were in the outcomes analysis middle age groups (35–54 years); only 3.9% After excluding from the schizophrenia were aged 18-24; 18.9% were aged 25–34, population all individuals whose diagnoses and 19.7% were in the oldest age group were only recorded in the last two years (55–64 years). or who had missing data on any of the matching variables (see section Methods), Health and social outcomes 15,639 individuals with schizophrenia Table 3 shows a range of outcomes for the remained. A pool of 2,691,342 individuals schizophrenia population and the matched without a schizophrenia diagnosis existed comparison group for 2015. Individuals in from which to select the matched cohort. the schizophrenia group were significantly more likely than the comparison group to Table 2 displays the demographic charac- have a general hospital admission (OR 1.18), teristics of the analytical sample (N=31,278) a psychiatric inpatient admission (OR 52.80) after one-to-one matching. Note that the or an emergency department attendance one-to-one matching resulted in the schizo- (OR 2.35). Individuals with schizophrenia

Table 1: Prevalence of schizophrenia in New Zealand as at 31 December 2015 by age, sex and ethnicity.a

Number with Prevalence (per b schizophrenia 1,000) OR (95% CI) Overall 18,096 6.7

Sex Male (reference group) 11,577 8.9 -

Female 6,519 4.7 0.52 (0.51, 0.54)

Ethnicityc European 11,730 6.3 0.77 (0.74, 0.79)

Māori 6,624 16.7 3.36 (3.26, 3.47)

Pacific 2,106 10.9 1.70 (1.62, 1.78)

Asian 1,065 3.0 0.39 (0.36, 0.41)

Age group 18–24 1,068 2.5 0.38 (0.35, 0.41)

25–34 (reference group) 3,582 6.5 -

35–44 4,854 8.9 1.37 (1.31, 1.43)

45–54 5,169 8.2 1.26 (1.21, 1.32)

55–64 3,423 6.3 0.96 (0.92, 1.01)

a Sub-group values may not sum to totals due to Stats NZ rounding requirements. b Ethnic comparisons are adjusted for age and sex. c Reference groups for ethnic comparisons are non-European, non-Māori, non-Pacific and non-Asian respectively.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 71 www.nzma.org.nz/journal article

were significantly more likely to be justice sector for those with and without receiving a social welfare benefit (OR 17.64) schizophrenia. Costs were significantly and living in social housing (OR 3.39). Justice higher for the schizophrenia group than in system interactions were also significantly the matched comparison group across each more common among the group with schizo- of these domains. The largest difference phrenia; they were around twice as likely was in health costs, where the average to have police proceedings initiated against per-person cost for the schizophrenia group them (OR 2.15), receive a conviction (OR was nearly seven times higher than for the 1.78) and receive a prison sentence (OR 2.49) control group. Social support costs were than individuals in the comparison group. five times higher among the schizophrenia Finally, individuals with schizophrenia group and justice sector costs nearly two were significantly less likely to be employed times higher. than the matched comparison group (OR 0.11), and while their odds of receiving any Discussion income were higher than the control group This paper identified a national cohort of (OR 2.40), their average annual income was working age individuals living with schizo- less than half that of those without schizophrenia in New Zealand, examined a range of phrenia. People with schizophrenia were health and social outcomes for those indi- also significantly more likely to live in areas viduals and estimated associated government of higher socioeconomic deprivation (mean costs using information available from linked NZDep2013 score 7.4 compared to 5.9). administrative data in the IDI. Government costs The schizophrenia population included Table 4 displays the estimated government 18,096 individuals corresponding to a preva- costs for health, social support and the

Table 2: Demographic characteristics of the analytical sample as at 31 December 2015 by age, sex and ethnicity.*

N %

Overall 31,278

Sex Male 19,959 63.8

Female 11,322 36.2

Ethnicity European 20,445 65.4

Māori 11,403 36.5

Pacific 3,504 11.2

Asian 1,743 5.6

Age group 18–24 1,209 3.9

25–34 5,901 18.9

35–44 8,646 27.6

45–54 9,363 29.9

55–64 6,162 19.7

* Sub-group values may not sum to totals due to Stats NZ rounding requirements.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 72 www.nzma.org.nz/journal article

Table 3: Health, social support, justice sector and economic outcomes for schizophrenia and matched comparison populations.a

Matched Schizophre- non-schizo- nia popula- phrenia OR / difference in means tion population (95% CI) (N=15,639) (N=15,639)

Health Non-mental health inpatient stay (%) 13.0 11.2 1.18 (1.10, 1.26)

Mental health inpatient stay (%) 12.2 0.3 52.80 (38.72, 72.01)

Emergency department visit (%) 6.7 3.0 2.35 (2.10, 2.62)

ACC compensation (%) 16.3 29.8 0.46 (0.43, 0.48)

Social support Any social welfare benefit (%) 78.8 17.4 17.64 (16.68, 18.67)

Social housing (%) 13.2 4.3 3.39 (3.09, 3.71)

Justice sector Police proceedings against (%) 9.4 4.6 2.15 (1.96, 2.36)

Conviction (%) 4.7 2.7 1.78 (1.58, 2.01)

Prison sentence (%) 2.3 0.9 2.49 (2.05, 3.03)

Economic Employed (%) 27.6 78.2 0.11 (0.10, 0.11)

Any income (%) 91.0 80.7 2.40 (2.24, 2.57)

Mean income ($) 17,865 44,092 -26,226 (-27,067, -25,384)

Mean NZDep13 (decile) 7.4 5.9 1.53 (1.47, 1.59)

a Sub-group values may not sum to totals due to Stats NZ rounding requirements.

Table 4: Government health, social support and justice sector costs (in New Zealand dollars) for schizophrenia and matched comparison populations.a

Schizophrenia Matched non-schizo- population phrenia population Cost difference (95% CI) (N=15,639) (N=15,639) Health ($) 17,333 2,486 14,847 (14,340, 15,354)

Social support ($) 14,667 2,845 11,823 (11,649, 11,997)

Justice sector ($) 795 431 363 (267, 459)

Total government ($) 32,795 5,762 27,033 (26,467, 27,599)

a Costs may not sum to totals due to Stats NZ rounding requirements.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 73 www.nzma.org.nz/journal article

lence of 6.7 per 1,000 people in the working phrenia have higher rates of interactions age (18 to 64) population. This estimate is with the criminal justice system than the in line with previous international esti- general population aligns with existing mates of the prevalence of schizophrenia,7,8 literature.39–42 However, our finding that but it is slightly higher than previous New approximately one in ten working-age indi- Zealand estimates, a difference that may viduals with schizophrenia have criminal be accounted for by the previous estimates justice system involvement is at the lower having used smaller time windows32 or end of existing estimates.41,42 Methodological more limited data sources.19 The higher differences likely explain some of this prevalence for Māori reported in this study variation. However, it is also possible that is consistent with previous New Zealand some of this difference is due to the time studies that used administrative data and period covered: the risk of offending may be found schizophrenia incidence and preva- highest in the pre-diagnosis period,43 which lence rates around two to three times higher our study did not include. 19,32 for Māori than for non-Māori. Higher Our estimated employment rate among rates among Māori could reflect higher the schizophrenia group of 27.6% is rates of schizophrenia. But these rates could consistent with findings from a recent inter- also reflect that Māori are more likely than national review where employment rates non-Māori to be given a schizophrenia diag- ranged from 11.9% to 39.0%.14 Our finding 33 nosis when they have psychosis symptoms. of low incomes among the schizophrenia Our finding that the prevalence of schizo- population, in particular relative to the phrenia is higher among males than females non-schizophrenia group, is also consistent adds to the existing literature where conclu- with existing literature.15,44 This is likely sions about sex difference in prevalence are explained by low employment rates and the 8,34 inconclusive. Prevalence rates are highest high proportion of individuals receiving among the 35–54 age group. The lower social welfare benefits that provide a low rate for the 55–64-year-old cohort is likely income. International evidence has demon- because our available data do not enable strated that employment among individuals us to look back far enough to the peak with schizophrenia can decrease the (early adult) diagnosis ages for that group. reliance on mental health services.45,46 Given In contrast, for the younger cohorts, in the substantial use of healthcare asso- particular those aged 18 to 24, it is probable ciated with schizophrenia, this suggests that that a schizophrenia diagnosis has yet to be targeting resources towards providing stable formalised for many. and meaningful employment for those with Compared with matched controls, people schizophrenia may be cost saving. This is living with schizophrenia had more ED an area of research that could be explored visits and more inpatient hospital admis- further. sions for both mental and physical health This is the first study to provide infor- reasons. The high rates of mental health mation about the government costs for a service use are unsurprising given the New Zealand schizophrenia cohort, which substantial mental health burden of living totalled $32,795 per person and were more with schizophrenia, and increased ED than five times higher than for individuals and hospital use has been reported previ- without schizophrenia, consistent with 35,36 ously. There is evidence that people living overseas studies.13,17,47,48 The largest areas of with schizophrenia have poorer physical costs, both in absolute terms and relative health and higher rates of co-occurring to the non-schizophrenia population, were 11,37,38 medical conditions, and this, along with health ($17,333 per person, seven times ED visits directly related to schizophrenia, higher) and social support ($14,667 per may account for the high levels of ED and person, five time higher). hospital use. Our finding of lower ACC use Our reported costs for the schizo- among the schizophrenia population is phrenia population are lower than those likely due to lower employment rates in the in a 2010 Australian study that reported schizophrenia group and therefore reduced total government costs of AUD$55,403 eligibility for income compensation. per person. That study employed a preva- Our finding that people with schizo- lence-based, bottom-up approach utilising

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 74 www.nzma.org.nz/journal article

survey data and unit costs from government relevant population for measuring outcomes and non-government ogranisation (NGO) and costs. Linked health and non-health sources.18 Furthermore, the Australian study data enabled us to consider outcomes and included some large costs that were not government costs across a range of areas available for inclusion in our study, such including health, social support, justice and as primary healthcare and tax foregone. the labour market. Finally, we were able to The low employment rates and high use of select a comparison group of individuals healthcare among our schizophrenia popu- without schizophrenia, who were also living lation suggest that costs related to primary in New Zealand, and match these to the care and tax foregone would be high and, if schizophrenia population on age, sex and included, would bring our estimates closer ethnicity. From this we are able to under- to those of the Australian study. stand how outcomes and costs for those with The cost estimates in this paper were schizophrenia compare to an equivalent drawn directly from recorded costs in the subset of the general population without IDI without additional refinements, and schizophrenia. therefore may contain some level of error. The use of administrative data also has In addition there are government costs that some limitations. The identification of were not able to be captured from the IDI, schizophrenia diagnoses relied on records of such as the costs of primary and private hospital admissions and specialist secondary healthcare, which may have led to an under- mental health service use. Although we estimation of the true government cost of expect that most individuals with diagnosed schizophrenia. However, as the same costing schizophrenia will have been treated by methods were applied to the schizophrenia specialist mental health services at some and comparison groups, the relative differ- point, we will have missed a number of ences in costs between the two groups are individuals who were being managed solely likely to be more reliable. in primary care, by NGOs or by family and Social sector cost data have not been whānau. A New Zealand study indicated widely used for research and there is little as many as 30% of those with psychosis do 50 information available about the best way to not seek medical help at all. In addition, use these costs. One exception is the health our method of identification relies on the sector, where complex methods do exist accuracy of diagnosis attribution as well for refining health costs.49 It may be useful as the quality of clinical coding. The extent to develop similar protocols for using costs to which these result in incorrect identifi- from other sectors, such as justice or social cations or missed identifications remains support. Incorporating these refinements unknown. would improve the accuracy of the cost Although the comparison group in this estimates in future work that uses cost data study was matched to the schizophrenia from the IDI. population on age, sex and ethnicity, it is Schizophrenia has many costs that cannot possible that the comparison population be captured from existing administrative differs from the schizophrenia population data. Examples include caregiver costs, in other ways that may be related to the time and productivity loss and intangible measured outcomes, such as deprivation costs related to pain and suffering. Detailed or family background. Due to the restricted survey data, such as those used in the time series of many datasets in the IDI, for Australian psychosis costing study,18 may be most people these potential matching vari- required to quantify those costs that are not ables were not available prior to the onset collected in administrative data. of schizophrenia. Future studies may be able to match more comprehensively as the The use of linked administrative data from time series of data in the IDI continues to the IDI confers a number of advantages. lengthen. Whole-of-population data meant we were able to identify a national cohort living with Schizophrenia is one of the most prospec- schizophrenia in New Zealand. In addition, tively consistent psychosis diagnoses over 51,52 individuals were removed from the analysis time. However, we acknowledge that if they died or moved overseas after their for a small proportion of our schizophrenia schizophrenia diagnosis, leaving a more cohort there may be a recovery or a change

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 75 www.nzma.org.nz/journal article

in diagnosis over time. Further research Administrative data are a useful tool for could explore the extent to which different providing a whole-population perspective diagnostic trajectories are associated with on the impact of mental health conditions. different health, social and economic This is especially true when multiple data outcomes. sources are combined in resources such as The use of administrative data in New the IDI, which allow us to consider impacts Zealand for research purposes is legal, across different domains. Further method- and the use of IDI data is highly regulated. ological development, especially around Nonetheless, there are ethical concerns, costing, would improve the usefulness of including concerns around social licence this resource for researchers. and especially in relation to the IDI, which Disclaimer now links data at the individual level These results are not official statistics. across a range of domains. As discussed They have been created for research by Bowden et al, further consultation purposes from the Integrated Data Infra- around these concerns is necessary, in structure (IDI) which is carefully managed particular given that comparisons across by Stats NZ. For more information about the ethnic groups may disadvantage Māori and IDI please visit https://www.stats.govt.nz/ Pasifika.53,54 integrated-data/.

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 76 www.nzma.org.nz/journal article

Appendix

Appendix Figure 1: diagnosis codes used to define schizophrenia • Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) codes of 295.1 to 295.9, excluding 295.7, or • International Classification of Diseases, Tenth Edition, Australian Modification (ICD-10-AM) codes of F20.0 to F20.9, or • International Classification of Diseases, Ninth Edition, (ICD-9) diagnosis codes of 295.0 to 295.9, excluding 295.7.

Appendix Table 1: Health, social support, justice sector and economic outcomes for schizophrenia and matched comparison populations for (five-year schizophrenia diagnosis).a

Schizophrenia Matched OR / difference in population non-schizo- means (95% CI) (N=12,045) phrenia population (N=12,045)

Health Non-mental health inpatient stay (%) 13.0 10.6 1.25 (1.16, 1.36)

Mental health inpatient stay (%) 11.5 0.2 67.69 (44.79, 102.29)

Emergency department visit (%) 6.7 2.7 2.56 (2.25, 2.92)

ACC compensation (%) 16.8 29.5 0.48 (0.45, 0.51)

Social support Any social welfare benefit (%) 78.0 17.4 16.83 (15.79, 17.94)

Social housing (%) 13.2 4.3 3.35 (3.02, 3.71)

Justice sector Police proceedings against (%) 8.6 4.2 2.15 (1.92, 2.39)

Conviction (%) 4.1 2.8 1.45 (1.26, 1.67)

Prison sentence (%) 1.9 0.9 2.04 (1.62, 2.56)

Economic Employed (%) 27.5 78.0 0.11 (0.10, 0.11)

Any income (%) 91.0 80.4 2.45 (2.27, 2.65)

Mean income ($) 18,175 43,754 -25,575 (-26,477 -24,673)

Mean NZDep13 (decile) 7.4 5.8 1.58 (1.51, 1.65)

a Sub-group values may not sum to totals due to Stats NZ rounding requirements.

Appendix Table 2: Government health, social support and justice sector costs (in New Zealand dollars) for schizophrenia and matched comparison populations (five-year schizophrenia diagnosis).a

Schizophrenia Matched non-schizo- population phrenia population Cost difference (95% CI) (N=12,045) (N=12,045) Health ($) 16,695 2,619 14,070(13,491, 14,649)

Social support ($) 14,760 2,907 11,850 (11,649, 12,054)

Justice sector ($) 642 414 228 (129, 324)

Total government ($) 32,096 5,942 26,150 (25,505, 26,795)

a Costs may not sum to totals due to Stats NZ rounding requirements.

Appendix Table 3: Health, social support, justice sector and economic outcomes for schizophrenia and matched comparison populations (five matched controls for every schizophrenia case).a

Schizophrenia Matched non-schizo- population phrenia population OR / difference in means (N=15,639) (N=78,195) (95% CI)

Health Non-mental health inpatient stay (%) 13.0 10.7 1.24 (1.18, 1.31)

Mental health inpatient stay (%) 12.2 0.2 56.12 (48.35, 65.15)

Emergency department visit (%) 6.7 2.9 2.45 (2.27, 2.64)

ACC compensation (%) 16.3 29.7 0.46 (0.44, 0.48)

Social support Any social welfare benefit (%) 78.8 17.5 17.58 (16.85, 18.34)

Social housing (%) 13.2 4.3 3.37 (3.18, 3.57)

Justice sector Police proceedings against (%) 9.4 4.5 2.19 (2.05, 2.33)

Conviction (%) 4.7 2.6 1.82 (1.67, 1.99)

Prison sentence (%) 2.3 0.9 2.56 (2.26, 2.91)

Economic Employed (%) 27.6 78.2 0.11 (0.10, 0.11)

Any income (%) 90.9 80.5 2.43 (2.3, 2.57)

Mean income ($) 17,865 44,092 -26,226 (-27,067, -25,384)

Mean NZDep13 (decile) 7.4 5.8 1.57 (1.52, 1.62)

a Sub-group values may not sum to totals due to Stats NZ rounding requirements.

Appendix Table 4: Government health, social support and justice sector costs (in New Zealand dollars) for schizophrenia and matched comparison populations (five matched controls for every schizophrenia case).a

Schizophrenia Matched non-schizo- population phrenia population Cost difference (95% CI) (N=15,639) (N=78,195) Health ($) 17,331 2,535 14,802 (14,535, 15,072)

Social support ($) 14,667 2,868 11,796 (11673, 11,916)

Justice sector ($) 795 411 387 (321, 453)

Total government ($) 32,798 5,815 26,983 (26,668, 27,300)

a Costs may not sum to totals due to Stats NZ rounding requirements

Appendix Table 5: Health, social support, justice sector and economic outcomes for schizophrenia and matched comparison populations (10 matched controls for every schizophrenia case).a

Schizophrenia Matched non-schizo- population phrenia population OR / difference in means (N=15,639) (N=156,390) (95% CI)

Health Non-mental health inpatient stay (%) 13.0 11.2 1.18 (1.10, 1.26)

Mental health inpatient stay (%) 12.2 0.3 52.80 (38.72, 72.01)

Emergency department visit (%) 6.7 3.0 2.35 (2.10, 2.62)

ACC compensation (%) 16.3 29.7 0.46 (0.44, 0.48)

Social support Any social welfare benefit (%) 78.8 17.4 17.64 (16.68, 18.67)

Social housing (%) 13.2 4.3 3.39 (3.09, 3.71)

Justice sector Police proceedings against (%) 9.4 4.6 2.15 (1.96, 2.36)

Conviction (%) 4.7 2.7 1.78 (1.58, 2.01)

Prison sentence (%) 2.3 0.9 2.49 (2.05, 3.03)

Economic Employed (%) 27.6 78.2 0.11 (0.10, 0.11)

Any income (%) 91.0 80.7 2.4 (2.24, 2.57)

Mean income ($) 17,865 44,092 -26,226 (-27,067, -25,384)

Mean NZDep13 (decile) 7.4 5.9 1.53 (1.47, 1.59)

a Sub-group values may not sum to totals due to Stats NZ rounding requirements.

Appendix Table 6: Government health, social support and justice sector costs (in New Zealand Dollars) for schizophrenia and matched comparison populations (10 matched controls for every schizophrenia case)a

Schizophrenia Matched non-schizo- population phrenia population Cost difference (95% CI) (N=15,639) (N=15,639) Health ($) 17,333 2,486 14,862 (14,649, 15,075)

Social support ($) 14,667 2,845 11,769 (11,655, 11,889)

Justice sector ($) 798 435 363 (300, 423)

Total government ($) 32,795 5,762 26,983 (26,668, 27,300)

a Costs may not sum to totals due to Stats NZ rounding requirements.

Competing interests: Mr Bowden, Dr Brewer and Dr Gibb reports grants and personal fees from Janssen Cilag Pty Ltd during the conduct of the study. Acknowledgements: This research was funded by grants provided by Janssen Cilag Pty Ltd to the University of Otago. Author information: Sheree Gibb: Department of Public Health, University of Otago, Wellington, New Zealand. Naomi Brewer: Department of Public Health, University of Otago, Wellington, New Zealand; Centre for Public Health Research, College of Health, Massey University, Wellington, New Zealand. Nicholas Bowden: Department of Women’s and Children’s Health, University of Otago, Dunedin, New Zealand. Corresponding author: Sheree Gibb: Department of Public Health, University of Otago, Wellington, New Zealand, PO Box 7343, Wellington South, Wellington, New Zealand 6242 [email protected] URL: www.nzma.org.nz/journal-articles/antipsychotic-and-sedative-medication-use-in-long-term-care-facilities-providing-dementia-care

REFERENCES 1. Vos T, Barber RM, Bell B, 4. Jääskeläinen E, Juola P, 8. Saha S, Chant D, Welham Bertozzi-Villa A, Biryukov Hirvonen N, McGrath JJ, J, McGrath J. A systematic S, Bolliger I, et al. Global, Saha S, Isohanni M, et review of the prevalence regional, and national al. A Systematic Review of schizophrenia. PLoS incidence, prevalence, and and Meta-Analysis of medicine. 2005;2(5):e141. years lived with disability Recovery in Schizophrenia. 9. Hjorthøj C, Stürup AE, for 301 acute and chronic Schizophrenia Bulletin. McGrath JJ, Nordentoft diseases and injuries in 2012;39(6):1296-306. M. Years of potential life 188 countries, 1990–2013: a 5. van der Werf M, Hanssen lost and life expectancy in systematic analysis for the M, Köhler S, Verkaaik M, schizophrenia: a systematic Global Burden of Disease Verhey FR, van Winkel R, review and meta-analysis. Study 2013. The Lancet. et al. Systematic review The Lancet Psychiatry. 2015;386(9995):743-800. and collaborative recalcu- 2017;4(4):295-301. 2. Mason P, Harrison G, lation of 133 693 incident 10. Hor K, Taylor M. Suicide Glazebrook C, Medley I, cases of schizophrenia. and schizophrenia: a Croudace T. The course Psychological Medicine. systematic review of rates of schizophrenia over 2014;44(1):9-16. and risk factors. Journal 13 years: a report from 6. Morgan VA, McGrath JJ, of psychopharmacology. the International Study Jablensky A, Badcock JC, 2010;24(4_suppl):81-90. on Schizophrenia (ISoS) Waterreus A, Bush R, et 11. Cunningham R, Sarfati coordinated by the World al. Psychosis prevalence D, Peterson D, Stanley Health Organization. The and physical, metabolic J, Collings S. Premature British Journal of Psychi- and cognitive co-morbid- mortality in adults using atry. 1996;169(5):580-6. ity: data from the second New Zealand psychiatric 3. Galletly C, Castle D, Australian national services. N Z Med J. Dark F, Humberstone V, survey of psychosis. 2014;127(1394):31-41. Jablensky A, Killackey Psychological Medicine. 12. Knapp M, Mangalore E, et al. Royal Australian 2014;44(10):2163-76. R, Simon J. The global and New Zealand College 7. Simeone JC, Ward AJ, Rotel- costs of schizophrenia. of Psychiatrists clinical la P, Collins J, Windisch R. Schizophrenia bulletin. practice guidelines for the An evaluation of variation 2004;30(2):279-93. management of schizophre- in published estimates of 13. Mangalore R, Knapp nia and related disorders. schizophrenia prevalence M. Cost of schizophre- Australian & New Zealand from 1990─2013: a system- nia in England. The Journal of Psychiatry. atic literature review. BMC journal of mental health 2016;50(5):410-72. Psychiatry. 2015;15(1):193.

policy and economics. 2008;42(11):941-9. Crampton P. NZDep2013 2007;10(1):23-41. 21. Wheeler AJ. Treatment index of deprivation. 2014. 14. Jonsdottir A, Waghorn pathway and patterns 29. SAS Institute Inc. SAS G. Psychiatric disor- of clozapine prescribing 7.1. Cary, NC: SAS ders and labour force for schizophrenia in Institute Inc; 2014. activity. Mental health New Zealand. Annals 30. R Core Team. R: A language review journal. 2015. of Pharmacotherapy. and environment for statis- 15. Hakulinen C, McGrath JJ, 2008;42(6):852-60. tical computing. Vienna, Timmerman A, Skipper N, 22. Joyce PR. Changing trends Austria: R Foundation for Mortensen PB, Pedersen in first admissions and Statistical Computing; 2019. CB, et al. The association readmissions for mania 31. StataCorp. Stata Statis- between early-onset schizo- and schizophrenia in New tical Software: Release phrenia with employment, Zealand, 1974 to 1984. 15. College Station, TX: income, education, and Australian & New Zealand StataCorp LLC.; 2017. cohabitation status: nation- Journal of Psychiatry. 32. Kake T, Arnold R, Ellis P. wide study with 35 years of 1987;21(1):82-6. Estimating the prevalence follow-up. Social Psychiatry 23. Gureje O, Miles W, Keks of schizophrenia among and Psychiatric Epidemiol- N, Grainger D, Lambert New Zealand Māori: ogy. 2019;54(11):1343-51. T, McGrath J, et al. Olan- a capture–recapture 16. Chong HY, Teoh SL, Wu zapine vs risperidone approach. Australian DB-C, Kotirum S, Chiou in the management of and New Zealand C-F, Chaiyakunapruk N. schizophrenia: a random- Journal of Psychiatry. Global economic burden ized double-blind trial in 2008;42(11):941-9. of schizophrenia: a Australia and New Zealand. 33. Petrović-van der Deen FS, systematic review. Neuro- Schizophrenia research. Cunningham R, Manuel J, psychiatric disease and 2003;61(2-3):303-14. Gibb S, Porter RJ, Pitama S, treatment. 2016;12:357. 24. Milne BJ, Atkinson J, et al. Exploring indigenous 17. Cloutier M, Aigbogun MS, Blakely T, Day H, Douwes ethnic inequities in first Guerin A, Nitulescu R, J, Gibb S, et al. Data episode psychosis in New Ramanakumar AV, Kamat Resource Profile: The Zealand–A national cohort SA, et al. The economic New Zealand Integrated study. Schizophrenia burden of schizophrenia Data Infrastructure (IDI). Research. 2020;223:311-8. in the United States in International journal of 34. McGrath J, Saha S, Welham 2013. The Journal of epidemiology. 2019. J, El Saadi O, MacCauley clinical psychiatry. 25. Zhao J, Gibb S, Jackson C, Chant D. A systematic 2016;77(6):764-71. R, Mehta S, Exeter DJ. review of the incidence 18. Neil AL, Carr VJ, Constructing whole of of schizophrenia: the Mihalopoulos C, Mackinnon population cohorts for distribution of rates and A, Morgan VA. Costs of health and social research the influence of sex, psychosis in 2010: Findings using the New Zealand urbanicity, migrant status from the second Australian Integrated Data Infrastruc- and methodology. BMC National Survey of Psycho- ture. Australian and New medicine. 2004;2(1):13. sis. Australian & New Zealand Journal of Public 35. Salsberry PJ, Chipps E, Zealand Journal of Psychi- Health, in press. 2017. Kennedy C. Use of general atry. 2014;48(2):169-82. 26. Gibb S, Bycroft C, Mathe- medical services among 19. Tapsell R, Hallett C, Mellsop son-Dunning N. Identifying Medicaid patients with G. The rate of mental health the New Zealand resident severe and persistent service use in New Zealand population in the Inte- mental illness. Psychiatric as analysed by ethnicity. grated Data Infrastructure Services. 2005;56(4):458-62. Australasian Psychiatry. (IDI). Statistics New 36. Hackman AL, Goldberg 2018;26(3):290-3. Zealand; 2016. Report RW, Brown CH, Fang LJ, No.: 0908350333. 20. Kake TR, Arnold R, Ellis Dickerson FB, Wohlheiter P. Estimating the preva- 27. Statistics New Zealand. K, et al. Use of emergency lence of schizophrenia IDI Data Dictionary: department services among new Zealand M a Programme for the integra- for somatic reasons ori: a capture–recapture tion of mental health data by people with serious approach. Australian October 2015 edition. 2015. mental illness. Psychiatric and New Zealand 28. Atkinson J, Salmond C, services. 2006;57(4):563-6. Journal of Psychiatry.

37. De Hert M, Correll CU, Survey on Alcohol and burden of schizophrenia Bobes J, Cetkovich-Bakmas Related Conditions Wave in Canada in 2004. Current M, Cohen D, Asai I, et al. III. The Journal of clinical medical research and opin- Physical illness in patients psychiatry. 2019;80(2). ion. 2005;21(12):2017-28. with severe mental disor- 43. Silverstein SM, Del Pozzo 49. Blakely T, Foster R, Wilson ders. I. Prevalence, impact J, Roché M, Boyle D, N. Burden of Disease of medications and dispar- Miskimen T. Schizophrenia Epidemiology, Equity ities in health care. World and violence: realities and CostEffectiveness psychiatry. 2011;10(1):52. and recommendations. (BODE3) Study Protocol: 38. Jeste DV, Gladsjo JA, Crime psychology review. Version 1.02011. Avail- Lindamer LA, Lacro 2015;1(1):21-42. able from: https://www. JP. Medical comorbid- 44. Kessler RC, Heeringa S, otago.ac.nz/wellington/ ity in schizophrenia. Lakoma MD, Petukhova otago019355.pdf. Schizophrenia bulletin. M, Rupp AE, Schoenbaum 50. Gale C, Wells J, McGee 1996;22(3):413-30. M, et al. Individual and M, Oakley Browne M. A 39. Blaauw E, Roesch R, Kerk- societal effects of mental latent class analysis of hof A. Mental disorders in disorders on earnings in psychosis‐like experiences European prison systems: the United States: results in the New Zealand Mental Arrangements for mentally from the national comor- Health Survey. Acta disordered prisoners in bidity survey replication. Psychiatrica Scandinavica. the prison systems of American Journal of Psychi- 2011;124(3):205-13. 13 European countries. atry. 2008;165(6):703-11. 51. Heslin M, Lomas B, International Journal 45. Knapp M, Patel A, Lappin JM, Donoghue K, of Law and Psychiatry. Curran C, Latimer E, Reininghaus U, Onyeji- 2000;23(5-6):649-63. Catty J, Becker T, et al. aka A, et al. Diagnostic 40. Corrado RR, Cohen I, Hart Supported employment: change 10 years after a S, Roesch R. Compara- cost‐effectiveness across first episode of psychosis. tive examination of the six European sites. World Psychological Medicine. prevalence of mental Psychiatry. 2013;12(1):60-8. 2015;45(13):2757-69. disorders among jailed 46. Bush PW, Drake RE, Xie H, 52. Haahr U, Friis S, Larsen inmates in Canada and the McHugo GJ, Haslett WR. TK, Melle I, Johannessen United States. International The long-term impact of JO, Opjordsmoen S, et al. Journal of Law and Psychi- employment on mental First-episode psychosis: atry. 2000;23(5-6):633-47. health service use and diagnostic stability over 41. Swanson JW, Frisman LK, costs for persons with one and two years. Psycho- Robertson AG, Lin H-J, severe mental illness. pathology. 2008;41(5):322-9. Trestman RL, Shelton DA, Psychiatric Services. 53. Bowden N, Gibb S, Thabrew et al. Costs of criminal 2009;60(8):1024-31. H, Audas R, Camp J, Taylor justice involvement 47. Wu EQ, Birnbaum HG, B, et al. IDI trends in among persons with Shi L, Ball DE, Kessler antidepressant dispensing serious mental illness in RC, Moulis M, et al. The to New Zealand children Connecticut. Psychiatric economic burden of and young people between Services. 2013;64(7):630-7. schizophrenia in the United 2007/08 and 2015/16. The 42. Moore KE, Oberleitner LM, States in 2002. Journal New Zealand Medical Jour- Zonana HV, Buchanan AW, of Clinical Psychiatry. nal. 2019;132(1505):48-61. Pittman BP, Verplaetse TL, 2005;66(9):1122-9. 54. Durie M. Measuring Māori et al. Psychiatric disor- 48. Goeree R, Farahati F, Burke wellbeing. New Zealand ders and crime in the US N, Blackhouse G, O’Reilly D, Treasury Guest Lecture population: results from Pyne J, et al. The economic Series. 2006;1:2007-09. the National Epidemiologic

Cannabis Legalisation and Control Bill: should doctors be concerned? Guna Kanniah, Shailesh Kumar

ABSTRACT A referendum on the Cannabis Legalisation and Control Bill was held in New Zealand. The Bill was meant to oversee government control over the production, supply and use of cannabis and reduce cannabis-related harm. Public health control was proposed over cannabis market by imposing licenses and cultivation, the quality and strength of marketed cannabis, and sale restrictions. Under this Bill, cannabis was only meant to be available to adults aged over 20 years through licenced stores. The potency of cannabis was to be limited. Cannabis use and was going to be permitted in private homes and specifically licensed premises. The Electoral Commission announced on 6 November 2020 that 50.7% of voters opposed the Bill and 48.4% supported it. Despite the outcome of the referendum, legalisation of cannabis may remain a live issue for many people, and doctors need to have an informed view about the impact of legalisation on mental health conditions. Experience from other countries shows that access to and potency of cannabis increased with legalisation. Despite the intent to prevent harm, cannabis legislation has been associated with adverse effects on mental health, emergency hospital presentations and crime. Public health strategies, including educating public about harm associated with cannabis, surveillance of potency and labelling, increasing minimal age for legal recreational cannabis use and bolstering treatment capacity for problematic cannabis use, including those with psychiatric disorders, should be funded by revenue generated from cannabis legislation.

annabis is the most commonly used and doctors will need to have an informed illegal drug in New Zealand, even view on whether legalising cannabis can Cthough the unauthorised possession result in increased access, usage and harm of any amount of cannabis is a crime under to vulnerable sections of society. This article the Misuse of Drugs Act 1975. A non-binding examines data relevant to the health of New referendum was held on 17 October 2020 Zealanders who are potentially most at risk. regarding a proposed Cannabis Legalisation A recent paper has stated New Zealand 1 and Control Bill (the Bill). The Bill outlined has one of the highest rates of cannabis government control over the production, use in the Western world, and that rates supply and use of cannabis, with the intent nearly increased one-and-half times from to reduce cannabis-related harm to individ- 2011/2012 to 2016/2017.2 We also know from uals, families/whānau and communities. It two world-class New Zealand longitudinal did not cover medicinal cannabis, hemp, studies that a dose-dependent relationship driving while impaired or workplace health exists between cannabis use and a range and safety issues, which were already of adverse health outcomes, including loss 1 covered by existing laws. The Electoral of cognitive capacity, increased respiratory Commission of New Zealand released official and periodontal disease, poor educational results of the referendum on 6 November and occupational outcomes, higher rates 2020, with 50.7% of voters opposing the of criminal convictions, poor relationships legalisation and 48.4% in support. The and driving impairments.3 Furthermore, referendum is over but the issue of can- Māori are not only disproportionately nabis policy reform remains and deserves represented among the population using it,2 critical appraisal. Proponents of legalising but also among those facing legal compli- cannabis may revisit the issue in the future cations associated with cannabis usage.2,4

While prevalence of cannabis use in the to not allow sale of cannabis to adoles- general population for 2018/2019 was 15%, cents was positive because this group is at the figure for the same period jumped to higher risk of experiencing impairment in 32% for Māori.4 These trends are already psychological, social and/or occupational worrying, and any risk of their potential functioning and suffering negative psychi- worsening needs to be carefully assessed atric consequences.3 It is possible that while considering the appropriateness or taking a health-focused approach driven need for legalisation of cannabis. by legislation may encourage people with These statistics3,4 are based on the avail- complications associated with cannabis use ability of cannabis in the illegal market, to seek treatment and give them access to which could potentially change if cannabis regulated cannabis products without adul- is to be legalised. We acknowledge the vast terants. The exact impact of legalisation on majority of people who use cannabis in New help-seeking behaviour may not become Zealand do not suffer any serious health evident immediately. or social consequences. But the negative We do, however, know that a substantial effects on people who do suffer harm from proportion of people with mental illnesses cannabis is significant. Among people who in New Zealand already use cannabis and use cannabis, those with mental illnesses suffer adverse mental health outcomes.4 are particularly vulnerable. This paper There was no provision in the Bill1 to restrict primarily focusses on this subgroup of the people with existing mental illnesses from population. buying cannabis. With easier and increased Currently, the majority of cannabis users access to cannabis, the subgroup of the in New Zealand access cannabis from an population with existing mental illnesses, or uncontrolled, illicit market. Criminal and at risk of developing mental illnesses, will antisocial activities are associated with the use more cannabis and may suffer greater illicit cannabis market. Measures to legalise harm. The relationship between cannabis and control cannabis, as outlined in the use and harm to mental health is well docu- 5 Bill,1 would have put strict public-health mented, and yet activism and commercial controls over this market by imposing interests have contributed to legalisation licenses and regulation over the cultivation, of cannabis in many parts of the world. quality and strength of products made, and Murray and Hall (2020) recently noted, “the by restricting the sale of cannabis products. legalisation of cannabis production and Under the Bill, cannabis would have only sale has created a rapidly growing industry been available to adults aged 20 years or with a strong financial interest in promoting 6 older through specialist stores licensed by the cannabis use.” Emerging data from the USA Government. Cannabis would not have been show cannabis legalisation could indeed be sold to teenagers. The potency of cannabis linked with increased usage and increases was to be limited. Cannabis use was to be in motor vehicle accidents, alcohol use and 7 permitted in private homes and specifically incidents of overdose injuries. It remains licensed premises, and it could not have been unknown whether legalisation of cannabis, used in public. Marketing and advertising by affecting availability, access and restric- would have been banned to help prevent tions on usage, can truly disempower 2 and reduce youth use and consumer, and well-established black markets. health warnings would have been required. The prevailing belief behind the New In essence, the Government was to control Zealand referendum was that controlled the cannabis market from seed to sale and access to cannabis can be associated with proposed to use the revenue to fund mental better outcomes. It is highly likely that health and addiction services, freeing up legalisation would make access to cannabis resources currently used in fighting criminal easier for many New Zealanders, and that activities associated with cannabis.1 its overall use could increase along with It is our view legalising cannabis adverse effects on mental health. In this correlates with heightened acceptance, paper we review the experience of other reduced perception of risks and an countries who have legalised cannabis, increase in cannabis use in both adults in terms of its effect on access, usage and and adolescents. The proposal in the Bill1 mental health. We will not examine the

impact on mental health of New Zealand’s monthly use of cannabis in states that had medicinal cannabis scheme, which has been legalised cannabis increased compared in effect from 1 April 2020.8 to states where it was not legalised.11 Furthermore, cannabis was used in a Increased access to greater variety of forms, including concen- trate, vaped oils, edibles and drinks in states cannabis under a that had legalised it.11 Similarly, another legalised model American study reported legalisation of cannabis was associated with increased The Netherlands decriminalised recre- prevalence of cannabis-use disorder.7 It is ational cannabis in 1976 by permitting its highly likely that with legalisation and an sale in approved outlets and possession of availability of a wider range of products (eg, up to five grams for personal use. Contrary for smoking, in edible form etc) cannabis to expectations, changes in cannabis use use would increase in New Zealand, as has in Netherlands developed rather inde- occurred in other countries. pendently of cannabis policy.9 Korf (2002) examined nearly four decades of post-de- Most proponents of legalising cannabis criminalisation data and described two identify the benefits of reducing criminal waves of changes in cannabis use, which activities, minimising the harm associated first peaked around 1970, fell to a low during with cannabis use and protecting the youth, the late 1970s and early 1980s and then who are especially at a higher risk, from peaked again in the mid-1990s. Cannabis use the harmful effects of cannabis. The impact among youth in the Netherlands occurred of legalisation, however, varies on these in parallel to four identified stages in the parameters and data are still emerging. For availability of cannabis: peaking when instance, legalisation of cannabis created cannabis was distributed through an under- a stronger illicit market for cannabis sales 12 ground market (late 1960s and early 1970s), and associated criminal activity in Canada. decreasing when the number of house Early post-legalisation data from a subse- 13 dealers superseded the underground market quent nationally representative study in the post-decriminalisation period, but of cannabis use and related behaviours, again stabilised or slightly decreased by the conducted in the months immediately end of the 1990s when the number of coffee before and after cannabis was legalised in shops was reduced. The Dutch experience, Canada, indicated that cannabis use among therefore, shows questionable effects of youth had not increased. Cannabis use in changes in cannabis policy on trends in the older age group increased in the short cannabis use. Indeed, it has been ques- and longer term post legalisation. Driving tioned whether legalisation and easy access after using cannabis did not change post to cannabis can reduce the overall usage of legalisation. The survey acknowledged cannabis in a given country.2 post-legalisation users had continued to procure cannabis from, and share it with, In other countries that have legalised family and friends. The overall risk of cannabis, the prevalence of cannabis use has developing cannabis use disorder post legal- increased over time along with associated isation increased in Canada.13 The findings complications. For example, in the USA of this survey suggest that, while criminal legalisation was followed by increased emer- activities associated with cannabis may gency department visits for cannabis-related reduce with legalisation, other associated presentations (cannabis intoxication and harms may not. cannabis-related hyperemesis).10 In New Zealand, the Bill1 was criticised for being Furthermore, the concentration of tetra- vague about the cannabis production and hydrocannabinol (THC) in cannabis has supply chain, despite proposing that a wide steadily increased with legalisation, from range of cannabis products will be available, approximately 3% in many traditional including raw (fresh and dried) cannabis, herbal forms to anywhere between 10% 14,15 resin, cannabis concentrates and cannabis and 70% in Europe and North America. infused products such as edibles and drink- The wider availability of cannabis in the 29 ables.2 A study from Northern America states of the USA that had ‘Medical Mari- found the prevalence of daily, weekly and juana Laws’ was associated with increased

cannabis potency between 1990 and 2014, risk of psychosis has increased.6 Increasing more unintentional childhood exposures access to cannabis, especially to potent between 2000 and 2013 and greater adult cannabis, may increase the risk of devel- cannabis use and adult cannabis use oping psychosis, particularly in the younger disorder between 2002 and 2014.15 These age group. The Dunedin Multidisciplinary changes were reported over a period of two Health and Development Study found decades, from 1990 to 2010.15 This trend age-related associations between cannabis is worrying because use of high-potency use and mental disorder. Mental disorder cannabis is associated with increased risk at age 15 led to a small but significantly of developing psychotic disorders. A multi- elevated risk of cannabis use at age 18. By centre case control study spread across contrast, cannabis use at age 18 elevated Europe and Brazil found that if high-potency the risk of mental disorder at age 21. These cannabis were no longer available, 12.2% findings suggest that the primary causal of cases of first-episode psychosis could be direction leads from mental disorder to prevented across all the 11 sites.16 Placing cannabis use among adolescents and the a cap on potency can be helpful given the reverse in early adulthood17—findings harmful effects of potent cannabis. Expe- echoed by Meier et al.18 Despite the proposed rience from countries that have legalised controls around age in the Bill, people at cannabis does suggest potency increased greater risk of experiencing adverse mental over time.14,15 Canada has experienced health outcomes may still be vulnerable, significant policy changes post legalisation, although how legalisation will affect mental including legal sale of more potent products health parameters and usage of cannabis and edibles (both with their own associated may be difficult to predict. special risks) and the opening of the market The impact of cannabis use, especially for retail cannabis by removing the cap on use of high-potency varieties, on adverse the number of private stores in some states, mental health outcome is worth examining 13 like Ontario. This could also occur in New in greater detail. A cohort study (n= 1,087) Zealand even if a cap on potency was to be found that use of high-potency cannabis placed in any future law. was associated with a significant increase Thus legalisation in some countries has in the frequency of cannabis use, cannabis resulted in increased access to cannabis, problems and anxiety disorder. The like- in a diverse range of preparations and in lihood of psychotic experiences increased increased potency, especially in the highly among users of high-potency cannabis, but commercialised markets. The impact of the risk was attenuated after adjustment for any attempts to legalise and/or to control frequency of cannabis use.19 An Australian cannabis in New Zealand will have to have study, on the other hand, found cannabis provisions to rigorously monitor the potency use precipitated the onset of psychosis in the of cannabis post legalisation, given the vulnerable and exacerbated the course in relationship between use of potent cannabis people with existing psychosis.20 Individuals and the risk of adverse outcomes, including with psychosis who are regular cannabis escalation to other drug use, especially in users have more positive symptoms, more the younger population.3 frequent relapses and require more hospital- isations.21 Regular cannabis use predicts an Cannabis use and increased risk of schizophrenia, even after controlling for confounding variables.21 In mental illnesses a meta-analysis, the pooled estimate for the If it is true that with legalisation access time course between regular cannabis use to and use of cannabis increases, then any initiation and age at onset of psychosis was negative effects on mental illnesses need to 6.3 years.22 This meta-analysis challenged be considered carefully. The relationship the popular notion that cannabis is initiated between cannabis use and the risk of by many as a form of self-medication for the developing psychotic symptoms has been positive symptoms of psychosis, although well documented.3 In countries that have cannabis may have some anxiolytic effects.22 legalised or decriminalised cannabis, its With such a well-established relationship price has fallen while dependence and the between cannabis use and risk of developing

or aggravating psychosis, the potential of to prepare a framework for clinicians and increased harm among people with existing, policy-makers to approach these concerns or at risk of developing, mental illnesses by incorporating the following steps:26 needs to be considered if cannabis were to i. A sound general population education become widely available with no safeguards strategy to limit access to cannabis for people in the ii. Limits on cannabis potency and high-risk category. clearer product labelling It is not just the risk of psychosis that iii. A minimum age for legal recreational increases with cannabis use. A range of cannabis use potential harms in patients with psychotic and mood disorders secondary to cannabis iv. A national surveillance strategy have been increasingly documented in other before and after cannabis legalisation countries.5 A greater level of depressive strategy symptoms are also reported in heavy v. Developing an enhanced treatment cannabis users compared to light users capacity for problematic cannabis and nonusers.23 Associations between use, such as for those with psychiatric cannabis use and negative outcomes in disorders bipolar affective disorder, such as worsened There was provision in the Bill1 for an affective episodes, psychotic symptoms, education strategy to be implemented. Here rapid cycling, suicide attempts, decreased we emphasise that legalising drugs like long-term remission, poorer global func- cannabis needs to occur in conjunction with tioning and increased disability, have been evidenced-based preventive and early inter- reported.5,24 After controlling for multiple vention efforts to reduce harmful cannabis confounders, cannabis use predicted use, with a strong focus on education.3 the development of anxiety disorders, Taking a public-education approach was, depression, suicidal ideation (nearly therefore, a positive aspect of the Bill, along threefold), personality disorders and inter- with the provisions to control potency personal violence, especially in adolescents and enforce clear product labelling and relative to adults, and a younger age of initi- a minimum age of 20 for recreational ation increases the risk of developing mental cannabis use—by introducing this age health disorders.25 limit, we stood the chance of reducing the In summary, cannabis is known to dispro- harm caused by early adolescent initiation, portionately harm people who are either although young adults may still have been at risk of mental illnesses or who have an at risk of experiencing adverse mental existing mental illness.5 Current data from health conditions, particularly psychosis. England, Denmark and Portugal indicate The choice of 20 years as the minimum age the incidence of schizophrenia and hospi- was criticised as “being mainly founded talisation rates for psychotic conditions in opinion or speculation combined with increased post cannabis legalisation.6 Such political calculations rather than concrete risks may increase with greater access scientific evidence”.2 Furthermore, New to cannabis, especially if potency was to Zealand policy-makers need to consider increase. Doctors as a group, and psychia- the impact of legalisation on the health and trists in particular, will need to proactively wellbeing of people who are younger than monitor and make submissions on how 20 years old.2 people at risk of mental illness, or with In Canada, the proportion of people existing mental illnesses, are affected by the accessing cannabis from friends and legalisation of cannabis in New Zealand. We family and thus not paying for cannabis may achieve greater clarity in this regard as did not change with legalisation.13 This legalisation and liberalisation of cannabis suggests at least some young people in New occurs in many countries, including New Zealand may continue to access cannabis Zealand, and as we accumulate empirical from friends and family post legalisation. data that will help us understand the role Taking a gradual, educational approach, of public policy on cannabis legalisation.12 rather than holding a binary referendum, Until then, we can use the experiences of has been proposed as better alternatives other countries that have legalised cannabis to legalisation.3 Post-legalisation creep in

activism for more potent cannabis and for wider range of cannabis products, such as Conclusions raw (fresh and dried), resin, vaping concen- The data presented above are relevant trate and edibles, to be made available6 for regulators, public health officials and has occurred in other parts of the world.6,11 policymakers considering the impact of This could also occur in New Zealand. An legalisation of cannabis for recreational effective system of surveillance for limiting use as in New Zealand. These findings potency and enforcing labelling across also have implications for mental health the full range of products will be essential policy in terms of education on risks and because, compared to dried cannabis, harm-minimisation strategies for products other products have different properties, containing cannabis, and for research into such as delayed onset of effect and higher effects in people who might be vulnerable to potency.11 Consumers will need to be mental illness. There are strong reasons to informed of such differences. If possible, approach cannabis legalisation cautiously. the legislative framework should protect We need to closely monitor the impact of people with mental illnesses from the different forms of legalisation of cannabis in harms associated with increased access to other countries as we evaluate the effects of cannabis. changes in our own.

Competing interests: Nil. Author information: Guna Kanniah: M.Clin.Pharm., PG Cert.Psychopharmacotherapy, Senior Clinical Pharmacist, Mental Health and Addictions Services, Waikato Hospital, PO Box 3200 Hamilton, New Zealand Shailesh Kumar: FRANZCP, MRCPsych, MPhil (London), DPM, Dip CBT, MD (Auck), Consultant Psychiatrist, Midland Regional Forensic Psychiatric Service, Honorary Clinical Associate Professor, University of Auckland Corresponding author: Guna Kanniah M.Clin.Pharm., PG Cert.Psychopharmacotherapy, Senior Clinical Pharmacist, Mental Health and Addictions Services, Pharmacy Services, Waikato Hospital, Selwyn Street, Hamilton 3204, +6421 54 99 27 [email protected] URL: www.nzma.org.nz/journal-articles/cannabis-legalisation-should-doctors-be-concerned

REFERENCES 1. Wilkins C, Rychert M. 3. Poulton R, Robertson K, explore-topics. Assessing New Zealand’s Boden J, Horwood J, Theo- 5. Lowe D, Sasiadek J, Coles Cannabis Legalization dore R, Potiki T, Ambler A. A, George T. Cannabis and Control Bill: prospects Patterns of recreational and mental illness: and challenges. Addiction. cannabis use in Aotearoa- a review. European 2021 Feb;116:222-230. New Zealand and their Archives of Psychiatry doi: 10.1111/add.15144. consequences: evidence to and Clinical Neuroscience. Epub 2020 Jul 4. inform voters in the 2020 2018 April; 269:1-14. referendum. Journal of 2. Fischer B, Bullen C. 6. Murray RM, Hall W. Will the Royal Society of New Emerging prospects for non Legalisation and Commer- Zealand. 2020. 50:2. 348-65. medical cannabis legal- cialization of Cannabis doi: 10.1080/03036 isation in New Zealand: Use Increase the Incidence 758.2020.1750435 An initial view and and Prevalence of Psycho- contextualization. Inter- 4. Ministry of Health [Inter- sis? JAMA Psychiatry. national Journal of Drug net]. Available from: 2020; 77(8):777-8. Policy. 2020. Feb;76:102632. https://minhealthnz. 7. Delling FN, Vittinghoff doi: 10.1016/j. shinyapps.io/nz-health-sur- E, Dewland TA, Pletcher drugpo.2019.102632. vey-2018-19-annual-data- MJ, Olgin JE, Nah G, Epub 2019 Dec 24. explorer/_w_c21841e3/#!/

Aschbacher K, Fang CD, International Perspectives S, Zammit S, Hickman M, Lee ES, Fan SM, Kazi DS, on the Implications of Cannon M, et al. Asso- Marcus GM. Does canna- Cannabis Legalisation: ciation of High-Potency bis legalisation change A Systematic Review & Cannabis Use With Mental healthcare utilisation? A Thematic Analysis. Int Health and Substance population-based study J Environ Res Public Use in Adolescence. JAMA using the healthcare cost Health. 2019;(16):3095. psychiatry. 2020;(e201035). and utilisation project in 13. Rotermann M. What has 20. Degenhardt L, Hall W, Lyns- Colorado, USA. BMJ Open. changed since cannabis key M. Testing hypotheses 2019 May 15;9(5):e027432. was legalized? Health Rep. about the relationship doi: 10.1136/bmjop- 2020 Feb 19;31(2):11-20. between cannabis use and en-2018-027432. PMID: doi: 10.25318/82 psychosis. Drug Alcohol 31092662; PMCID: -003-x202000200002- Depend. 2003; 71(1):37-48. PMC6530411. eng. PMID: 32073644. 21. Hall W, Degenhardt L. 8. Medsafe [Internet]. 14. Hall W, Stjepanović D, Cannabis use and the risk Medicinal Cannabis Caulkins J, Lynskey M, of developing a psychotic Scheme. Prescriber Update. Leung J, Campbell G, et disorder. World Psychi- 2020 [cited 2020 August al. Public health impli- atry. 2008; 7(2):68-71. 10]. Available from: cations of legalising the 22. Myles H, Myles N, Large https://www.medsafe. production and sale of M. Cannabis use in govt.nz/profs/PUArticles/ cannabis for medicinal and first episode psychosis: PDF/Prescriber_Update_ recreational use. Lancet. Meta-analysis of preva- Vol_41(1)_M arch_2020.pdf. (2019); 394(10208):1580-90. lence, and the time course 9. Korf DJ. Dutch coffee shops 15. Hasin D. US epidemiol- of initiation and continued and trends in cannabis ogy of cannabis use and use. Aust N Z J Psychiatry. use. Addict Behav. 2002 associated problems. 2016; 50(3):208-19. Nov-Dec;27(6):851-66. Neuropsychopharmacol- 23. Lev-Ran S, Roerecke M, Le doi: 10.1016/s0306- ogy. 2018; 43(1):195-212. Foll B, George T, McKenzie 4603(02)00291-5. 16. Di Forti M, et al. The K, Rehm J. The associa- PMID: 12369472. contribution of cannabis tion between cannabis 10. Leung J, Chiu J, Chan V, use to variation in the use and depression: a Chan G, Stjepanović D, incidence of psychotic systematic review and Hall W. What have been disorder across Europe meta-analysis of longi- the public health impacts (EU-GEI): a multicentre tudinal studies. Psychol of cannabis legalisation case-control study. Lancet Med.. 2014; 44(4):797–810. in the USA? A review of Psychiatry. 2019; 6(427–36). 24. van Rossum, I, et al Does evidence on adverse and 17. McGee R, Williams S, Cannabis Use Affect Treat- beneficial effects. Current Poulton R, Moffitt T. A ment Outcome in Bipolar Addiction Reports. 2019; longitudinal study of Disorder? A Longitudinal 6(4): p. 4) 418-428. cannabis use and mental Analysis J Nerv Ment 11. Goodman S, Wadsworth health from adolescence to Dis 2009;197:35–40 E, Leos-Toro C, Hammond early adulthood. Addiction. 25. Copeland J, Rooke S, Swift D; International Canna- 2000; 95; 95(4):491-503. W. Changes in cannabis bis Policy Study team. 18. Meier M, Hill M, Small P, use among young people: Prevalence and forms of Luthar S. Associations of impact on mental health. cannabis use in legal vs. adolescent cannabis use Curr Opin Psychiatry. illegal recreational canna- with academic perfor- 2013; 26(4):325-29. bis markets. Int J Drug mance and mental health: Policy. 2020 Feb;76:102658. 26. George T, Hill K, Vaccarino A longitudinal study of doi: 10.1016/j. F. Cannabis Legalisation upper middle class youth.. drugpo.2019.102658. and Psychiatric Disorders: Drug Alcohol Dependence. Epub 2020 Jan 9. Caveat “Hemp-tor”. The 2015; 156:207-12. PMID: 31927413. Canadian Journal of Psychi- 19. Hines L, Freeman T, Gage atry. 2018; 63(7):447-50. 12. Bahji A, Stephenson C.

Zoledronate-induced anterior uveitis, scleritis and optic neuritis: a case report Laura E Wolpert, Andrew R Watts

isphosphonates, such as zoledronate, with proptosis, periorbital oedema, conjunc- are used by approximately 55,000 tival chemosis and injection and cells and Bpeople per year in New Zealand1 to flare in the anterior chamber (Figure 1A). prevent the loss of bone density in a range There was no evidence of vitritis, and fundal of conditions such as osteoporosis, Paget’s examination was normal. 2 disease of the bone and bone metastases. A B-scan of the right eye showed scleral Although ocular side effects are rare, bis- thickening (Figure 2). The patient underwent phosphonates have been associated with a CT scan of her orbits, which revealed right- 3,4 4,5 acute anterior uveitis (AAU) and scleritis. sided proptosis with intraconal fat stranding There have also been case reports of optic and inflammation surrounding the globe 6–8 neuritis following bisphosphonate use. and optic nerve, consistent with scleritis and Here we report a case of a patient who pro- retrobulbar optic neuritis (Figure 3). Inves- gressively developed AAU, scleritis and optic tigations, including serum ACE, treponemal neuritis following a zoledronate infusion. serology, ANA and QuantiFERON-TB Gold, were unremarkable. Case report A diagnosis of zoledronate-induced uveitis, A 61-year-old woman with a past medical scleritis and optic neuritis was made. The history of previous morbid obesity with patient received 1g intravenous methyl- sleeve gastrectomy, severe reflux and prednisolone, which resulted in a rapid ileostomy secondary to hemicolectomy for improvement of her symptoms and signs by severe diverticular disease presented to the following day (Figure 1B). The patient the eye clinic with a three-day history of was then discharged on a weaning course right eye pain, photophobia and blurred of oral prednisone, topical prednisolone 1% vision. These symptoms commenced one eye drops and cyclopentolate 1% eye drops. day following her first zoledronate infusion At one week follow-up the inflammation had for osteoporosis. She had no significant past resolved. ophthalmic history. Right eye visual acuity was 6/15 and Discussion left eye visual acuity was 6/9. Intraocular Orbital inflammation is an uncommon pressures were normal. She had right side effect of zoledronate infusion. The circumlimbal injection with cells and flare incidence of zoledronate-associated AAU in the anterior chamber. Fundal exam- has been reported at around 1.1%.3 To our ination was normal. An initial diagnosis of knowledge, there are only two case reports AAU was made and treatment with prednis- of zoledronate-associated optic neuritis,7,8 olone 1% eye drops and cyclopentolate 1% although optic neuritis has been seen with eye drops was commenced. other bisphosphonates in a few cases.6 Two days later the patient presented There is little information on bisphos- with worsening pain and vision and pain phonate rechallenge following adverse on eye movements. Right visual acuity had ocular events. Adverse ocular events have decreased to 6/24. She had red desaturation been reported following bisphosphonate

Figure 1: (A) The patient’s right eye five days following a zoledronate infusion, showing proptosis, lid oedema and conjunctival chemosis. (B) The patient’s right eye after treatment with intravenous methyl- prednisolone, which resulted in reduced peri-ocular swelling and chemosis. The pupil is dilated due to cyclopentolate drops.

Figure 2: A B-scan of the patient’s right eye showing scleral thickening (white arrow).

rechallenge but do not occur in all cases.9,5–7 Recognition of drug-induced ocular Although inflammation associated with inflammation is critical to allow for prompt bisphosphonate use is usually mild and referral to an ophthalmologist and with- shows complete resolution after cessation drawal of the drug in question. Patients of the precipitating agent and treatment of receiving bisphosphonate treatment should the ocular inflammation, in the context of be counselled to seek medical attention if potentially sight-threatening conditions such they develop symptoms of visual loss, eye as scleritis and optic neuritis, rechallenge pain or eye redness. may not be advisable.

Figure 3: CT scan showing right eye proptosis (blue arrow), intraconal fat stranding (yellow arrow) and optic nerve thickening (red arrow).

Competing interests: Nil. Author information: Laura E Wolpert: Ophthalmology Registrar, Department of Ophthalmology, Whangarei Hospital, Northland District Health Board, Maunu Road, Whangārei 0148, New Zealand. Andrew R Watts: Consultant Ophthalmologist, Department of Ophthalmology, Whangarei Hospital, Northland District Health Board, Maunu Road, Whangārei 0148, New Zealand. Corresponding author: Dr Laura Wolpert, Department of Ophthalmology, Whangarei Hospital, Northland District Health Board, Maunu Road, Whangārei 0148, New Zealand, 09 430 4100 [email protected] URL: www.nzma.org.nz/journal-articles/zoledronate-induced-anterior-uveitis-scleritis-and-optic-neuritis-a-case-report

REFERENCES 1. Best Practice Advocacy 4. Keren S, Leibovitch Clin Oncol (R Coll Radiol). Centre New Zealand I, Ben Cnaan R, et al. May 2013;25(5):328- [Internet]. An update on Aminobisphosphonate-as- 9. doi:10.1016/j. bisphosphonates. Accessed sociated orbital and ocular clon.2012.12.006 Jan, 2021. Available from: inflammatory disease. 8. Lavado FM, Prieto MP, www.bpac.org.nz Acta Ophthalmol. Aug Osorio MRR, Gálvez MIL, 2. Drake MT, Clarke BL, 2019;97(5):e792-e799. Leal LM. Bilateral retrobul- Khosla S. Bisphospho- doi:10.1111/aos.14063 bar optic neuropathy as the nates: mechanism of 5. Samalia P, Sims J, Niederer only sign of zoledronic acid action and role in clinical R. Drug-induced ocular toxicity. J Clin Neurosci. Oct practice. Mayo Clin Proc. inflammation. N Z Med J. 2017;44:243-5. doi:10.1016/j. Sep 2008;83(9):1032-45. Dec 2020;133(1527):83-94. jocn.2017.06.048 doi:10.4065/83.9.1032 6. Stack R, Tarr K. 9. Patel DV, Horne A, Mihov B, 3. Patel DV, Bolland M, Drug-induced optic Stewart A, Reid IR, McGhee Nisa Z, et al. Incidence of neuritis and uveitis CN. The Effects of Re-chal- ocular side effects with secondary to bisphos- lenge in Patients with a intravenous zoledronate: phonates. N Z Med J. Mar History of Acute Anterior secondary analysis of a 2006;119(1230):U1888. Uveitis Following Intrave- randomized controlled 7. Brulinski P, Nikapota nous Zoledronate. Calcif trial. Osteoporos Int. AD. Zolendronic acid-in- Tissue Int. Jul 2015;97(1):58- Feb 2015;26(2):499- duced retrobulbar optic 61. doi:10.1007/ 503. doi:10.1007/ neuritis: a case report. s00223-015-0015-4 s00198-014-2872-5

Vildagliptin-induced bullous pemphigoid Annabelle Yi Zhang, Paul Jarrett

n New Zealand, vildagliptin became fully accompanying pruritus. Bullous pemphigoid subsidised by PHARMAC on 1 October is most commonly observed in the seventh I2018. Bullous pemphigoid is an auto- to ninth decade of life and is associated with immune blistering disorder that typically neurological diseases, including cerebrovas- affects the elderly. Dipeptidyl peptidase-four cular accidents.3 The mainstay of treatment inhibitors (DPP4is, or gliptins) are a cause of is steroid therapy. However, high-dose drug-induced bullous pemphigoid. systemic corticosteroid treatment (prednisolone equivalent >40mg daily) is associated Case presentation with significantly higher mortality during the first year.4 A randomised controlled A 69-year-old Caucasian man with type trial comparing the efficacy and safety two diabetes mellitus (T2DM) and no known profile of doxycycline and oral prednis- dermatological conditions presented with a olone concluded non-inferiority of high-dose one-month-old, widespread pruritic blis- doxycycline (200mg oral daily) as a first-line tering rash. This rash developed nine months treatment for bullous pemphigoid.1 after the addition of vildagliptin to his diabetic medication regimen. The Naranjo algorithm score of causality was five, indi- Figure 1: Bullous pemphigoid on the back, consist- cating a ‘probable’ adverse drug reaction. ing of vesicles, erosions from ruptured bullae and an erythematous urticated rash. There were widespread crusted erosions with scattered tense bullae over the scalp, trunk and limbs, and mucosal erosion on the lower lip (Figures 1 and 2). The diagnosis of bullous pemphigoid was confirmed by histology, which showed a subepidermal split with increased numbers of dermal eosinophils and positive anti-basement membrane antibody of 1:1280 titre. Direct immunofluorescence staining was not possible. Vildagliptin was ceased and oral doxycycline (200mg daily) initiated.1 Re-epithelialisation on doxycycline monotherapy progressed slowly, therefore oral prednisone (40mg daily) was added, resulting in rapid re-epithelialisation. Blood glucose levels were monitored closely and managed with metformin and correctional scale insulin while the patient was taking prednisone. Discussion Bullous pemphigoid is an autoimmune condition. Autoantibodies target hemides- mosomes in basal keratinocytes, causing loss of adhesion between the epidermis and dermis.2 It is characterised by localised or generalised bullae with preceding and/or

Figure 2: Mucosal involvement. exact mechanism remains unknown.3 Five cases of bullous pemphigoid in patients who had been on dual metformin and DPP4i therapy for 2 to 13 months prior to disease onset were described, and two cases were resistant to immunosuppressive therapy but later achieved stable remission upon cessation of DPP4is.8 A meta-analysis of case-control studies further supported this association and found that vildagliptin had a higher degree of association with bullous pemphigoid compared to sitagliptin and linagliptin.9 The other available oral hypo- glycaemic agents available in New Zealand do not cause blistering as a common adverse effect. DPP4is have become a popular second-line In New Zealand, vildagliptin became fully treatment in T2DM as they do not cause subsidised by PHARMAC on 1 October 2018. weight gain and have a lower adverse effect Before then, no cases of DPP4i-associated profile than sulphonylureas. However, in bullous pemphigoid had been reported to the last decade an increasing number of the Centre for Adverse Reactions Monitoring case reports and epidemiological studies (CARM). Eight cases were reported between have been published suggesting there is a October 2018 and September 2020.10 There relationship between bullous pemphigoid were six females, and the average age was and DPP4is.5-7 The latency time between the 75 years (standard deviation ± 10 years). initiation of DPP4is and onset of bullous The Australian Database of Adverse Event pemphigoid ranges from 1 to 37 months Notifications has recorded three cases of in case reports.5,6 The association between suspected vildagliptin-related pemphigoid. bullous pemphigoid and DPP4is was first This increased incidence reflects the wider described in 2011.8 There are several use of vildagliptin in the community since hypotheses about the pathogenesis of subsidisation. Bullous pemphigoid should be DPP4i-induced bullous pemphigoid, but the suspected in a patient on vildagliptin who develops an inflammatory rash with blisters.

Competing interests: Nil. Acknowledgements: Dr Natalie Poppito, Consultant Histopathologist, Middlemore Hospital. New Zealand Pharmacovigilance Centre, University of Otago Author information: Annabelle Yi Zhang, Medical Student: Faculty of Medical and Health Sciences, University of Auckland, New Zealand. Paul Jarrett, Dermatologist: Department of Dermatology, Middlemore Hospital, Counties Manukau District Health Board, Auckland, New Zealand; Department of Medicine, The University of Auckland, Auckland, New Zealand. Corresponding author: Dr Paul Jarrett, Module 7, Manukau SuperClinic, PO Box 98743, Manukau 2241, Auckland, New Zealand, +64 9 276 0000 (phone), +64 9 276 0282 (fax) [email protected] URL: www.nzma.org.nz/journal-articles/vildagliptin-induced-bullous-pemphigoid

REFERENCES 1. Williams H, Wojnarowska for Lethal Outcome in in Bullous Pemphigoid F, Kirtschig G, Mason J, Patients With Bullous Patients than in the French Godec T, Schmidt E et Pemphigoid. Archives of General Population. Journal al. Doxycycline versus Dermatology. 2002;138(7). of Investigative Dermatol- prednisolone as an initial 5. Béné J, Jacobsoone A, ogy. 2019;139(4):835-41. treatment strategy for Coupe P, Auffret M, Babai 8. Skandalis K, Spirova M, bullous pemphigoid: a S, Hillaire-Buys D et al. Gaitanis G, Tsartsarakis A, pragmatic, non-inferiority, Bullous pemphigoid Bassukas I. Drug-induced randomised controlled induced by vildagliptin: bullous pemphigoid in trial. The Lancet. a report of three cases. diabetes mellitus patients 2017;389(10079):1630-8. Fundamental & Clin- receiving dipeptidyl 2. Nishie W. Update on ical Pharmacology. peptidase-IV inhibitors the pathogenesis of 2014;29(1):112-4. plus metformin. Journal of bullous pemphigoid: An 6. Yoshiji S, Murakami the European Academy of autoantibody-mediated T, Harashima S, Ko R, Dermatology and Venere- blistering disease targeting Kashima R, Yabe D et ology. 2011;26(2):249-53. collagen XVII. Journal of al. Bullous pemphigoid 9. Phan K, Charlton O, Dermatological Science. associated with dipeptidyl Smith S. Dipeptidyl 2014;73(3):179-86. peptidase-4 inhibitors: A peptidase-4 inhibitors 3. Tasanen K, Varpuluoma report of five cases. Journal and bullous pemphigoid: O, Nishie W. Dipeptidyl of Diabetes Investiga- A systematic review and Peptidase-4 Inhibitor-Asso- tion. 2017;9(2):445-7. adjusted meta-analysis. ciated Bullous Pemphigoid. 7. Plaquevent M, Tétart F, Australasian Journal of Frontiers in Immunol- Fardet L, Ingen-Housz-Oro Dermatology. 2019;61(1). ogy. 2019;10. S, Valeyrie-Allanore L, 10. New Zealand Pharmacovig- 4. Rzany B, Partscht K, Jung Bernard P et al. Higher ilance Centre. Vildagliptin M, Kippes W, Mecking D, Frequency of Dipeptidyl Review. October 2020. Baima B et al. Risk Factors Peptidase-4 Inhibitor Intake

Endogenous endophthalmitis following computerised tomography colonography Deepesh Mehta, Helen Long, Neil Aburn

ndogenous endophthalmitis is a rare was given intravitreal ceftazadime and and potentially blinding condition vancomycin to both eyes prophylactically Eresulting from infection within the and started on intravenous acyclovir. The ocular tissues and accounts for 10% of all patient was found to be anaemic and to endophthalmitis cases.1 In contrast to exog- have an elevated erythrocyte sedimentation enous endophthalmitis, which typically re- rate and C-reactive protein. The remainder sults from external ocular procedures such of the blood tests were unremarkable. as cataract surgery, endogenous endoph- The vitreous PCR was negative for any thalmitis usually occurs in immunocompro- organisms and the TOE showed no evidence mised patients.1 Gram-positive species are of valvular vegetation. The blood culture the predominant causative organisms in was positive for Streptococcus dysgalactiae, bacterial endogenous endophthalmitis, with and the patient was subsequently started streptococcal and staphylococcal bacteria on intravenous ceftriaxone, to which he being the most commonly isolated species.1,2 responded well. He was discharged on oral amoxicillin, prednisone acetate eye Case report drops and dexamethasone eye ointment. At one-week follow-up, examination the BCVA A 64-year-old male presented to the acute improved to 6/6 and 6/9 in the right and left eye clinic with a one-day history of right eye respectively, with improving inflam- loss of vision, chills and muscle aches, after mation. At the final six-month follow-up, having undergone computerised tomog- his visual acuity was stable with no signs raphy colonography (CTC) two weeks prior of inflammation in either eye and he was for investigation of gastrointestinal (GI) discharged. bleeding while on anticoagulation. On examination, best corrected visual acuity (BCVA) in the right and left eye was 6/18 Discussion and 6/9 respectively. Additionally, there The pathophysiological mechanism of was evidence of bilateral anterior chamber endogenous endophthalmitis usually results inflammation, with keratic precipitates from bacteraemia secondary to seeding more severe in the right eye relative to the from other infected organs.1 Interestingly, left. Furthermore, a 0.2mm hypopyon was the right eye is more commonly affected evident in the right eye with mild vitritis. compared to the left, likely because it has a Right fundus examination showed solitary more direct arterial supply from the internal white-yellow fluffy lesions in the periphery carotid artery.1 and white-centred retinal haemorrhages. Streptococcus dysgalactiae is a type of A diagnosis of possible endogenous pyogenic beta haemolytic streptococci that endophthalmitis was made and the patient rarely causes endogenous endophthalmitis, underwent inflammatory blood tests, blood with reported cases being more common cultures, right vitreous tap, chest x-ray and in the elderly.3 It forms part of the natural trans-oesophageal echocardiogram (TOE) human microflora, occupying areas such to rule out infective endocarditis (IE). He as the skin and the gastrointestinal tract.3

Owing to its high virulence, infection of prophylaxis prior to CTC is not recom- ocular tissues can have devastating visual mended.4 Thus, colonic ischaemia secondary consequences.3 Infection of ocular tissue to CTC could have been the reason for devel- by this organism warrants investigation of opment of endogenous endophthalmitis in underlying IE.3 Additionally, infection of the present case.4 3 ocular tissues can cause secondary IE. In conclusion, endogenous endophthal- Patients undergoing CTC have a small mitis may present in patients following risk of acquiring transient bacteraemia CTC secondary to colonic insufflation. We secondary to colonic insufflation causing illustrate a novel case where Streptococcus local ischaemia, thereby allowing trans- dysgalactiae, which forms part of the normal location of enteric pathogens into the GI microflora, resulted in endogenous bloodstream and potential infection of endophthalmitis. Thus, clinicians should be distant tissues.4 However, the risk of bacte- aware of CTC as a rare cause of endogenous raemia is very low, and routine antibiotic endophthalmitis.

Competing interests: Nil. Author information: Deepesh Mehta: Ophthalmology Non-Training Registrar, Ophthalmology Department, Capital and Coast District Health Board, Wellington. Helen Long: Consultant Ophthalmologist, Ophthalmology Department, Capital and Coast District Health Board, Wellington. Neil Aburn: Consultant Ophthalmologist, Ophthalmology Department, Capital and Coast District Health Board, Wellington. Corresponding author: Deepesh Mehta, Ophthalmology Non-Training Registrar, Ophthalmology Department, Capital and Coast District Health Board, Wellington [email protected] URL: www.nzma.org.nz/journal-articles/endogenous-endophthalmitis-following-computerised-tomography-colonography

REFERENCES 1. Ness T, Pelz K, Hansen LL. case reports. 2018;2018. Society of Chemotherapy. Endogenous endophthal- 3. Hagiya H, Semba T, 2018;24(11):936-40. mitis: microorganisms, Morimoto T, Yamamoto 4. Ridge CA, Carter MR, disposition and prognosis. N, Yoshida H, Tomono K. Browne LP, Ryan R, Acta ophthalmologica Scan- Panophthalmitis caused by Hegarty C, Schaffer K, et dinavica. 2007;85(8):852-6. Streptococcus dysgalactiae al. CT colonography and 2. Mak CY, Sin HP, Chan subsp. equisimilis: A case transient bacteraemia: VC, Young A. Klebsiella report and literature implications for antibiotic endophthalmitis as review. Journal of infec- prophylaxis. European the herald of occult tion and chemotherapy : radiology. 2011;21(2):360-5. colorectal cancer. BMJ official journal of the Japan

The competition is next door! Why a voluntary approach to tobacco retailer reduction will never work Richard Portch

n 2011, Te Kāwanatanga o Aotearoa (The rather than the small area unit the address New Zealand Government) set the goal resides in. We calculated how many retailers Ito become smokefree by 2025 and have of the same type were within a 100m and less than 5% of the population smoking 250m radius of each retailer. 1 daily. A decade on, adult (15 years+) daily A total of 1,794 tobacco retailers and 425 smoking rates have decreased from 16.3% to community pharmacies were identified in the current 11.6%, over double the intended Tāmaki Makaurau. Across the three DHBs goal, and with 32% of wāhine Māori still there were an estimated 1,591,797 inhabi- 2 smoking daily, significant disparities still tants, with 144,144 (11.3%) of the adults who exist. Aotearoa is not on track to achieve responded to the New Zealand Census 2018 Smokefree 2025, and to do so will require a stating they smoke daily. Therefore, people significant reduction in uptake plus a quick have greater access to tobacco than phar- 3–5 increase in cessation. The reduction of macies, with one tobacco retailer for every tobacco supply is an important sector-sup- 887 inhabitants compared to one community ported strategy to ensure the success of pharmacy for every 3,745 inhabitants. This the Government’s Smokefree 2025 Action equates to one tobacco retailer for every 6–8 Plan. Marsh et al have previously pub- 80 people who smoke daily. Comparing the lished on the abundance of tobacco retailers, DHBs, there is one tobacco retailer for every identifying 5,243 tobacco retailers across 51 people who smoke in Auckland DHB, one 9 Aotearoa. They found that schools were for every 115 people who smoke in Counties more likely to have a tobacco retailer within Manukau DHB and one for every 86 people 500m or 1,000m as well as a greater number who smoke in Waitematā DHB. Tobacco of tobacco retailers compared to community retailers were similarly distributed across pharmacies. local boards and district health boards as In this letter we share preliminary community pharmacies, by percentage. findings from a tobacco retailer study across There was a significant difference in distri- the three Tāmaki Makaurau (Auckland) bution across council zones, with 25% of district health boards (DHBs) interrogating tobacco retailers being in residential zones tobacco accessibility and density, using versus 17.6% of community pharmacies. community pharmacies as a comparison. Tobacco retailers were found to be more This builds on the work of Marsh et al, densely located in proximity to each other using Census 2018 and New Zealand Index than pharmacies, with 55.4% of tobacco of Deprivation (NZDep2018) data. We have retailers having another tobacco retailer used a new methodology for attributing within 100m vs just 22.6% of pharmacies. At deprivation and demographics to an address 250m the difference was 75.9% and 51.1%, whereby the average of all Statistical Area respectively. An initial logistic regression 1 (SA1) within a 1km radius of the address model using retailer type as a predictor is calculated. The advantage of using this found tobacco retailers to have greater odds methodology is that the deprivation and of having a retailer within 100m and 250m demographics attributed to each address than pharmacies (OR: 4.26, 95%CI: 3.30–5.50, represent their surrounding neighbourhood,

p<0.001 and OR: 3.03, 95%CI: 2.45–3.76, The findings of our research demonstrate p<0.001, respectively). Including NZDep2018 how abundant tobacco retailers are across score and population with district health Tāmaki Makaurau and detail what could be board and council zone the association addressed in the Government’s Smokefree remained (OR: 5.75, 95%CI: 4.37–7.57, 2025 Action Plan. Tobacco retailers have p<0.001 and OR: 4.85, 95%CI: 3.76–6.26, previously expressed voluntarily stopping p<0.001, respectively). Next we compared to sell tobacco is not likely, particularly how many closely located retailers each type when other retailers close to them continue had and found 15% of tobacco retailers had to sell.10–12 The voluntary approach will not three or more other retailers within 100m, be enough to reduce tobacco supply and a whereas zero community pharmacies did. mandated reduction in supply is required, At 250m the difference was 42% and 13%, particularly in high deprivation areas, respectively. which are most affected by smoking.2

Competing interests: Richard Portch notes that they work for a health sector agency working in tobacco harm reduction and control. Author information: Richard Portch: Health Promoter/Data Analyst, Health Improvement Team, Auckland Regional Public Health Services, Auckland. Corresponding author: Richard Portch, Health Promoter/Data Analyst, Health Improvement Team, Auckland Regional Public Health Services, Cornwall Complex, Building 15 (Level 4), Greenlane Clinical Centre, Greenlane, Auckland [email protected] URL: www.nzma.org.nz/journal-articles/the-competition-is-next-door-why-a-voluntary-approach- to-tobacco-retailer-reduction-will-never-work

REFERENCES 1. New Zealand Parliament shinyapps.io/nz-health-sur- sales to only pharmacies [Internet]. Government vey-2019-20-annual-da- combined with cessation Final Response to Report ta-explorer/ (2019). advice: A modelling study of the Māori Affairs 3. NZIER [Internet]. N. Z. I. of the future smoking Committee on Inquiry of E. R. Smoking targets A prevalence, health and into the tobacco industry microsimulation analysis. cost impacts. Tob. Control in Aotearoa and the Available from: www. 28, 643–650 (2019). consequences of tobacco nzier.org.nz (2012). 7. Robertson, L & Marsh, L use for Māori, presented 4. Ikeda, T, Cobiac, L, Wilson, Estimating the effect of a to the House of Repre- N, Carter, K & Blakely, T. potential policy to restrict sentatives in accordance What will it take to get tobacco retail availability with Standing Order 248 to under 5% smoking in New Zealand. Tob. (J.1) - New. Available from: prevalence by 2025? Control 466–468 (2018) https://www.parliament. Modelling in a country doi:10.1136/tobacco- nz/en/pb/papers-presented/ with a smokefree goal. Tob. control-2018-054491. current-papers/docu- Control 24, 139–145 (2015). 8. Wilson, N., Thomson, G. ment/49DBHOH_PAP21175_1/ W., Edwards, R. & Blakely, government-final-re- 5. Tobias, M I, Cavana, R Y & T. Potential advantages sponse-to-report-of-the- Bloomfield, A. Application and disadvantages of an māori-affairs (2011). of a system dynamics model to inform invest- endgame strategy: A ‘sink- 2. Ministry of Health NZ ment in smoking cessation ing lid’ on tobacco supply. [Internet]. Annual Data services in New Zealand. Tob. Control 22, i18 (2013). Explorer 2018/19: New Am. J. Public Health 9. Marsh, L., Doscher, C., Zealand Health Survey 100, 1274–1281 (2010). Cameron, C., Robertson, L. [Data File]. Available & Petrović-van der Deen, F. from: https://minhealthnz. 6. Der Deen, F S P Van et al. Restricting tobacco S. How would the tobacco

retail landscape change R. Regulating the sale of dairy owners and manag- if tobacco was only sold tobacco in New Zealand: ers on tobacco retailing. through liquor stores, A qualitative analysis NZMJ 131, 1479 (2018). petrol stations or pharma- of retailers’ views and 12. Badu, E. & Fowler, E. ‘If cies? Aust. N. Z. J. Public implications for advo- government can’t stop the Health 1753-6405.12957 cacy. Int. J. Drug Policy smokes, no way we can stop (2020) doi:10.1111/1753- 26, 1222–1230 (2015). selling them’: why retailers 6405.12957. 11. Witt, M. et al. How want Government to act. 10. Robertson, L., Marsh, L., important to dairies is NZMJ 131, 1482 (2018). Hoek, J., McGee, R. & Egan, selling tobacco? Views of

Rupture of the Urinary Bladder 1921

.H., aged 65, arrived at Queenstown by trigone. All other organs healthy except boat at 7.30 p.m. on 27th April, 1921. the mucosa of the stomach, which showed JHe stayed at an hotel, where he was a signs of a small healed ulcer. No evidence stranger; had a good tea with a glass of beer. of poisoning. The pelvic peritoneum was Later obtained half a teaspoonful of sweet acutely inflamed and of a brilliant scarlet spirits of nitre from the proprietor, which colour, doubtless from the extravasated he took in a little gin, saying that his water urine. The prostate was enlarged, inflamed, was bad. At 1.30 the proprietor heard groans and contained pus. The bladder mucosa coming from his room, where he found him was injected and œdematous, and there was lying on the floor, evidently in very great a small uric acid stone the size of a bean agony, and asking for more nitre. Medical embedded in the bladder wall. Ureters were aid was sent for, but the patient had expired normal and not dilated, nor was the pelvis just prior to arrival. There was blood coming of the kidney. There were no signs of an from the penis. injury or external violence on the abdomen. At post-mortem a ruptured bladder Verdict given that death was due to shock was discovered. A ragged tear about one and collapse resulting from acute peritonitis inch long situated posteriorly above the caused by free urine in the abdomen.

URL: www.nzma.org.nz/journal-articles/rupture-of-the-urinary-bladder

Proceedings for the 256th Otago Medical School Research Society Masters/ Honours Student Speak Awards, Wednesday 4 November 2020

The light activated confirmed its expression and Microparticles chloride channel function in vitro. produced from human GtACR2 is a novel We successfully targeted papillomavirus type 16 potential therapeutic GtACR2 expression to the E6 and E7 expressing to treat chronic pain. cell body of iPSC-derived, keratinocytes regulate cultured human neurons. N Lyons1,2, L Schoderboeck1, This subcellular localisation is antigen-presentation MF Ibrahim2, SM Hughes1, required to influence chloride by Langerhans cells. 2 balance solely in the cellular RM Empson . V Ticar, B Nair, M Wilson, M Hibma. 1Department of Biochemistry, compartment where it is altered Department of Pathology, 2Department of Physiology, School in chronic pain. Whole-cell Dunedin School of Medicine, of Biomedical Sciences, Brain patch clamp electrophysiology University of Otago, Dunedin. Health Research Centre, Univer- of GtACR2-expressing neurons sity of Otago, Dunedin. showed light-induced ionic Persistent infection by human Chronic pain affects 1 in 5 photocurrents (N = 15 cells) and papillomavirus (HPV), such New Zealanders costing our transient silencing of action as HPV16, initiates around society approximately $14 potential activity. Illumination 5% of all human cancers. The billion in 2016. Opioids are used of GtACR2-expressing cultures progression from persistent to treat chronic pain but show successfully silenced neuronal infection to cancer is asso- poor long-term efficacy and network activity for the entire ciated with the over-expression high rates of addiction. Innova- 30 second illumination period, of HPV16 E6 and E7 (E6/ tions for effective, personalised as detected using a genetically E7), and modulation of anti- chronic pain treatments are encoded fluorescent calcium gen-presenting cells (APCs). desperately needed. Aberrant indicator (N = 3 wells, 21 Microparticles (MPs) are pain signalling, arising from recordings across 4 days). small (100 to 1000 nm), cell disinhibition of spinal cord pain membrane-derived vesicles. Our successful generation projection neurons, is proposed This study aimed to understand and in vitro validation of this to underlie chronic pain. the role of MPs produced from unique optogenetic tool, and Altered chloride transporter HPV16 E6/E7-expressing cells on its viral delivery vector, paves expression in these neurons antigen-presentation by Lang- the way for its in vivo testing disrupts their chloride balance erhans cells (LCs), the only APCs in an animal model of chronic eroding the effectiveness of that are co-located with HPV-in- pain. Our results provide strong inhibitory input. fected keratinocytes. evidence for the novel use of We propose that expressing GtACR2 to restore inhibition To measure the effects of and activating the light-ac- of spinal cord pain projection MPs on T cell proliferation tivated chloride channel neurons. GtACR2 is promising in response to LC-presented Guillardia theta anion channel as a specific, light-tuneable ther- ovalbumin (OVA), E6/E7 or rhodopsin 2 (GtACR2) in spinal apeutic to re-establish normal control (Ctrl) MPs purified from cord pain projection neurons pain signalling and alleviate E6/E7 expressing or control will restore their chloride chronic pain. C57Bl/6 mouse keratinocytes balance, re-establish inhibition, were incubated with bone Supported by a Maurice and and reduce pain. Here, we marrow-derived LCs, pulsed Phyllis Paykel Trust Honours produced a lentiviral vector with OVA and co-cultured with Scholarship in Medical and encoding a red fluorescence CD8 T cells. CD8 T cell prolifer- Health Sciences. tagged GtACR2 construct and ation was increased following

incubation with Ctrl MPs (76.23 relevant to drinking water, to the release of catecholamines ± 15.68, mean percent ± SD) affect CRC cell viability in vitro. from adrenal medullary chro- compared to no MPs (39.31 ± The effect of NDEA on CYP2E1 maffin cells into the circulation. 16.43, P < 0.01, Kruskal-Wallis, expression was also assessed This secretion is caused by a Dunn’s multiple comparisons). as a source of reactive oxygen rise in intracellular Ca2+ ([Ca2+] CD8 T cell proliferation was species which may serve as a i) in response to several phys- reduced following incubation mechanism of NDEA toxicity. iological stimuli. Increasing with E6/E7 MPs (48.80 ± 15) evidence suggests that the The cell viability of HT-29 compared to Ctrl MPs (N = 8 adrenal medullary anatomical and Caco-2 CRC cells following mice, Kruskal-Wallis, Dunn’s organisation and cell intercon- 72 hours exposure to nitrate multiple comparisons). When nectivity contributes greatly to (0 – 400 mg/L), nitrite (0 – 20 LCs derived from Trans- this [Ca2+]i. mg/L), or NDEA (0 – 200 nM) porter associated with Antigen was determined using the The current study thus aimed Processing 1 (TAP1) knockout sulforhodamine B assay (N = 3). to characterise this stress mice were used, CD8 T cell Western immunoblotting was response by examining the stim- proliferation was significantly used to assess the expression of ulus-dependent [Ca2+]i changes reduced in E6/E7 (22.84 ± 5.46, CYP2E1 at 24 and 72 hours in in individual chromaffin cells P < 0.001) and Ctrl (19.95 ± 7.47, both cell lines exposed to NDEA within rat adrenal slices. P < 0.0001) MP cultures (N = 8 (200 nM and 200 μM, N = 1). Adrenal vibratome sections mice, Two-way ANOVA, Sidak’s were prepared from adult male multiple comparisons). While no concentration of Sprague Dawley rats, loaded nitrate and nitrite significantly Here we show that MPs with the Ca2+-indicator Fluo-4, altered cell viability in either increase proliferation of CD8 and changes in [Ca2+]i recorded cell line, NDEA at 100 nM and T cells upon incubation with from individual cells using fluo- 200 nM significantly increased LC-pulsed OVA, and that E6/E7 rescence microscopy. cell viability in HT-29 cells MPs suppresses this effect. Addi- by 13% (± 3.18) and 12% (± Nicotinic stimulation, tionally, the effects of MPs were 3.79) respectively (mean ± SD, mimicking the acute stress ablated in the absence of TAP1 p < 0.05, ANOVA, Bonferroni response, induced a sharp [Ca2+] supporting the involvement adjustment), with no effect i rise which, in some cells (N = of the antigen-presentation observed in Caco-2 cells. Initial 17/66), returned to basal in the pathway of LCs. HPV16 E7 has Western immunoblotting continued presence of nicotine. been previously reported to indicates that NDEA at 200 nM In contrast to studies on isolated repress TAP1. The measurement and 200 μM increases CYP2E1 cells, most recorded cells (N of TAP1 expression in LCs could expression in both Caco-2 and = 49/66) entered a sustained, indicate a potential mechanism HT-29 cells at 24 hours, however elevated plateau phase for the for the E6/E7 MP suppression of two further biological replicates duration of ncotinic exposure. antigen-presentation. are required to confirm this. Pituitary adenylate cyclase-activating peptide (PACAP), Toxic NDEA as a product of An in vitro investigation mimicking a persistent stress drinking water nitrate may of the effect of response, increased [Ca2+] increase the cell viability of CRC i in approximately half of the environmental cells accounting for correla- recorded cells (N = 33/67), but contaminants tions between elevated drinking this response was delayed water nitrate and incidence of nitrate, nitrite, and by approximately 10 min CRC. However, toxicokinetic N-nitrosodiethylamine after exposure. Histamine, a studies would be required on colorectal cancer. non-neuronal chromaffin cell before extrapolating these in stimulator, induced a rapid MT Howes, RJ Rosengren. vitro results to true carcinogenic [Ca2+]i rise in most cells (N = risk from these environmental Pharmacology and Toxicology, 50/55). These responses varied contaminants. School of Biomedical Sciences, greatly between cells and University of Otago, Dunedin. consisted of multiple complex Epidemiological studies have Ca2+-imaging in peaks, differing in duration and correlated elevated concentra- rat adrenal slices magnitude. Despite intercel- tions drinking water nitrate reveals chromaffin lular differences, histaminergic with an increased risk of devel- cell heterogeneity. responses were highly repro- oping colorectal cancer (CRC). ducible within an individual Nitrate is rapidly absorbed W Aye, T Georgescu, S Bunn. cell. in vivo, where 0.02 – 0.04% Centre of Neuroendocrinolo- is converted to N-nitrosodi- Each examined stimulus gy, Department of Anatomy, produced heterogeneous ethylamine (NDEA), a class 2A School of Biomedical Sciences, responses with novel [Ca2+]i carcinogen. How such environ- University of Otago, Dunedin. mental contaminants affect CRC characteristics not reported in The ability to undergo phys- cells is not well understood. isolated chromaffin cells. The iological adaptations to stress This project investigated the results, therefore, provided is critical to life. An important potential for nitrate, nitrite, evidence for contributions aspect of this stress response is and NDEA at concentrations of cell-cell interactions to

Ca2+-signalling and subsequent Patients more likely to have exclusively activated by secretory output from the chro- seen a dietitian were: younger designer drugs), activated by maffin cells, while highlighting (P < 0.001); diagnosed with the synthetic DREADD agonist avenues for future research into Crohn’s disease (P = 0.001); clozapine-n-oxide (CNO, 2 mg/ cellular and functional hetero- had previous IBD surgery (P < kg), were used to selectively geneity of the adrenal medulla. 0.001); on biologic therapy (P = manipulate RFRP activity in 0.005). Common themes iden- the transgenic mice. Immu- Access to dietetic tified through general inductive nohistochemical staining services for analysis identified that there of the neuronal activity is a lack of publicly funded marker Fos-related antigen inflammatory bowel dietetic services, that dietitians confirmed downregulated disease patients in need specialist IBD knowledge, RFRP neuronal activity in New Zealand – a that patients/respondents want the inhibitory DREADD-ex- patient view. routine referrals to dietitians pressing mice (P < 0.0001) and and to have ongoing dietitian upregulated activity in the NE McCarthy1, M Schultz2, CL Wall1. access. excitatory DREADD-expressing 1Department of Medicine – mice (P < 0.0001) compared to Results indicate that there University of Otago, Christchurch non-DREADD controls in both 2 is inequitable and inadequate Campus, Department of Medicine sexes (t-tests). The effect on LH access to dietetic services - University of Otago, Dunedin pulse frequency (mean ± SEM) for IBD patients in NZ, with International guidelines was measured by tail tip blood variable referral rates and recommend that inflam- sample collection every 6 mins substantial numbers of patients matory bowel disease (IBD) for 3 hours in awake, free- required to pay for dietitian patients should have access to ly-moving mice. appointments. Poor access specialised dietitian support. increases malnutrition risk in As expected, control mice This patient group is at high risk this vulnerable patient group, had a marked reduction in LH of malnutrition and nutrition potentially leading to worse pulse frequency in response to interventions can reduce health outcomes and negative 4 days of subcutaneous ~100 mg disease activity and improve impacts on quality of life. glucocorticoid (corticosterone) quality of life. Anecdotal reports implant treatment (males: 1.38 suggest that New Zealand (NZ) Supported by the Rabbitt ± 0.19 before and 0.92 ± 0.06 IBD patient access to dietitians Fund, managed by the Friends pulses/h after corticosterone, P = is variable. of Dunstan Hospital and the 0.030, N = 8; females 1.54 ± 0.15 Department of Medicine, This research aimed to inves- pulse/h before and 0.38 ± 0.11 University of Otago, Dunedin tigate factors associated with pulses/h after corticosterone, P patient access to a dietitian and = 0.0001, N = 7). In contrast, LH whether access meets patients’ RFamide-related pulse frequency was rescued expectations. In early 2020, an peptide neurons from these glucocorticoid effects anonymous electronic survey modulate chronic in RFRP-silenced females (1.23 ± ± was disseminated to patients glucocorticoid- 0.10 before and 1.074 0.12 (and parents) by Crohn’s and pulses/h after corticosterone, Colitis NZ and IBD health induced reproductive P = 0.120, N = 9), but not males professionals via email and suppression. (1.29 ± 0.15 before and 0.86 social media, with a reach of ± 0.12 pulses/h after corticos- A Mamgain, GM Anderson. approximately 2000 people. terone, P = 0.030, N = 8) (2-way Quantitative responses were Department of Anatomy, ANOVA with corticosterone Centre of Neuroendocrinology, analysed via chi-square and treatment and RFRP inhibition University of Otago, Dunedin. Fisher’s exact tests and quali- as factors and Holm-Sidak’s tative responses were analysed Globally, one in four couples multiple comparison post-hoc using a general inductive are affected by infertility. test). Supporting a sexually-di- approach. Chronic stress is often a result, morphic role for RFRP neurons but also a possible cause of in LH suppression, upregu- The respondents (N = 407) reproductive dysfunction. lation of RFRP neuronal activity were mostly female (74%) and Regulation of gonadal function reduced LH pulse frequency NZ European (91%), 5% iden- by luteinizing hormone (LH) in female mice (controls: 1.7 ± tified as Māori. While 95% of from the pituitary gland is 0.01, N = 5; RFRP-excited: 0.6 ± respondents had topics they suppressed during chronic 0.19 pulse/h, N = 3; P = 0.004) would like to discuss with a stress. However, the mecha- but no effect was observed dietitian, only 52% had ever nisms are poorly understood. in male mice (controls: 1.1 ± seen a dietitian and 45% had We investigated whether 0.11, N = 5; RFRP-excited: 1.3 ± never been referred. Of those hypothalamic RFamide-re- 0.15 pulses/h, N = 5; P > 0.999) who had seen a dietitian, lated peptide (RFRP) neurons (t-tests). 37% had self-funded their mediate this suppressive effect appointment, many because of glucocorticoids. These results reveal a novel they were unable to access sex-specific requirement of publicly funded appointments. Cre recombinase-dependent RFRP neurons in modulating DREADDs (designer receptors

suppressive effects of stress 10 µM oCOm-21, respectively Recent discoveries show that steroids on LH secretion and compared to vehicle (P < 0.01). obesity-induced morphological highlight complexities in Moderately hypertrophic hearts changes to EAT are dissimilar neuronal signaling associated treated with 3 µM of oCOm-21 to changes in subcutaneous or with reproductive dysfunction. had a 1.96-fold decrease in visceral adipose tissues. While NLRP3 staining compared to the adipocyte size in subcutaneous, oCOm-21 has an vehicle (P < 0.05). No signifi- appendicular and pericardial anti-inflammatory cance was observed between adipose tissues increased in vehicle and 1 µM oCOm-21 relation to body mass index effect via the NLRP3 for both hypertrophic groups. (BMI), EAT adipocyte size did inflammasome. Western blotting indicated no not. Additionally, the well-es- significance between vehicle tablished relationship between FM Payne, S Thwaite, JC Harrison, and oCOm-21 treated hearts (P increased EAT thickness and IA Sammut. > 0.05). BMI was not observed in Māori, Department of Pharmacology nor Pacific people. Our study and Toxicology, School of These preliminary results aimed to investigate how Biomedical Sciences, Univer- strengthen the hypothesis that obesity affects EAT morphology sity of Otago, Dunedin. IRI evokes an inflammatory and EAT localization in response, resulting in cardiac During cardiac bypass proce- diverse ethnic New Zealand damage, which may be atten- dures, the heart is subjected populations. to repeat cycles of ischaemia uated with oCOm-21 via the reperfusion injury (IRI). Pro-in- inhibition of NLRP3. Although In post mortem cases of NZ/ flammatory signalling involving the immunohistochemical European (N = 57), Māori (N = the NLRP3 inflammasome, results demonstrates a prom- 16) and Pacific (N = 6) people, has been identified as a key ising link between CO and adipocyte size was determined contributor in the pathogenesis NLRP3 inhibition, the lack of in histological slices of different of cardiac remodelling. Recent reproducibility with Western adipose depots, whereas research highlighted the capa- blotting requires additional myocardial adipose infiltration bility of carbon monoxide (CO) research. The current research and fibrosis were determined releasing compounds to reduce performed has provided in histological slices of the left injury via the NLRP3 inflam- evidence that CO can reduce ventricle. In subcutaneous, masome. This study aimed to myocardial IRI through the appendicular and pericardial test whether a novel organic CO NLRP3 inflammasome within a adipose tissue, the adipocyte releasing molecule, oCOm-21, diseased heart model similar to size positively correlated with produces an anti-inflammatory healthy hearts. BMI in NZ/European and Māori effect by reducing NLRP3 levels populations, while not in Pacific within hearts undergoing IRI. Epicardial adipose cases. Additionally, there was tissue morphology no correlation between EAT Male CYP1a1-Ren2 rats were adipocyte size and BMI in any induced to develop either diversity in Māori, cases (P = 0.706). Furthermore, non-hypertrophic or moder- Pacific and New EAT adipocyte size positively ately hypertrophic hearts before Zealand/European correlated with myocardial being randomly allocated to post-mortem cases. fibrosis in NZ/European P( = vehicle or oCOm-21 (1 – 10 0.0015, r2 = 0.5263) however not µM) treatment groups prior GRR Hight1, HM Aitken-Buck1, IC in Māori/Pacific (P = 0.2954). to IRI (N = 3/group). Hearts Fomison-Nurse1, S Coffey2, RD Tse3, Finally, a positive correlation were sectioned for histology RR Lamberts1. was observed between EAT to evaluate the damage from 1Department of Physiology, School adipocyte size and myocardial IRI and immunofluorescence of Biomedical Sciences, University adipocyte infiltration, but only staining for determination of of Otago, Dunedin. 2Department in Māori/Pacific (P = 0.023, r2 = NLRP3 expression. Ventricles of Medicine, HeartOtago, Dune- 0.5455). were homogenised for Western din School of Medicine, Dunedin blotting to determine protein Hospital, Dunedin, New Zealand. In conclusion, the differ- expression levels. Results were 3Department of Forensic Pathology, ences in adipocyte morphology analysed using a One-way LabPLUS, Auckland City Hospital, between Māori, Pacific people, ANOVA with a post-hoc Auckland, New Zealand. and NZ/European, highlight Bonferroni analysis. Obesity is the leading cause possible physiological ethnic of morbidity and heart disease variances, which may associate Untreated controls in both in New Zealand (NZ), and Māori with diverse risk factors and normotrophic and hypertrophic and Pacific people are dispro- disease characteristics. This groups sustained greater histo- portionately affected. Obesity is should trigger discussions on logical damage than oCOm-21 characterized by adipose tissue the validity of current diag- treated hearts. Immunofluo- expansion. The fat surrounding nostic and potential treatment rescence examination showed the heart, Epicardial Adipose strategies for cardiovascular a 2- and a 3-fold decrease in tissue (EAT), has recently been disease and obesity in NZ NLRP3 within non-hypertrophic identified as a diagnostic tool populations. hearts treated with 3 µM and for cardiovascular disease.

Characterisation of of inflammation in theex vivo (I) tissue at T1 compared to the inflammatory model that has the potential T-1, indicating that an inflam- to reduce animal usage and matory response to injury is response to injury in suffering. occurring (N = 2). In addition, an ex vivo rodent model IL-1β was upregulated by 1.9- Spinal cord segments of spinal cord injury. and 6.5-fold, for UI and I tissue were dissected from adult at T1 compared to T-1, again CBM McCrostie1, L Wise2. Sprague-Dawley rats (7 indicating that an inflammatory 1 weeks post-natal), with Department of Pharmacology response to injury is occurring random segments receiving a and Toxicology, School of (N = 2). Also, at T1, increased Biomedical Sciences, Univer- compression injury after 24 hr activation of macrophages sity of Otago, Dunedin. culture (T0). Segments were through calprotectin staining harvested immediately after Spinal cord injury (SCI) is a was observed while increased dissection (T-1), or 24 hr (T1) significant injury, and a burden expression of resident microglia and 7 days (T7) after injury. to the patient, their family and through mannose staining was Quantitative PCR and immuno- to the healthcare system. The observed. inflammatory response that histochemistry were performed follows the injury exasperates to assess the mRNA levels of These results suggest that tissue damage and scarring, inflammatory regulators and pro-inflammatory signalling by which prevents neuronal regen- the distribution of resident resident immune cells occurs eration and patient recovery. immune cells, respectively. within the ex vivo model of Current in vivo models are SCI at 24 hr post-injury. This Interleukin (IL)-6 expression problematic due to the extreme model may therefore have was upregulated by 1400- and suffering of animals receiving utility for testing therapeutics 4200-fold, respectively, for the injury. This study aimed to aimed at reducing SCI-induced uninjured (UI) and injured assess the extent and source inflammation.

URL: www.nzma.org.nz/journal-articles/proceedings-for-the-256th-otago-medical-school-research- society-masters-honours-student-speak-awards-wednesday-4-november-2020

Abstracts for the 255th Otago Medical School Research Society PhD Student Speaker Awards, Wednesday 19 August 2020

Theoretical reduction Home-Care (interRAI-HC), to 2Department of Pathology, Otago Medical School – Dunedin Campus, in the anticholinergic determine the reduction in the ACB. Using a Paired-Samples University of Otago, Dunedin. burden in older Test, we compared the results Cortical malformations arise adults with dementia of the ACB before and after the from in utero disruption of in New Zealand theoretical intervention of the neurogenesis, the process of pharmacological alternatives proliferation, differentiation, SS Bala1, HA Jamieson2, to MAP. and migration of neurons in PS Nishtala3, R Braund1. the developing brain. Finding 1 The 2015 interRAI dataset Department of Preventive and attributable genetic causes constituted 75,410 commu- Social Medicine, Dunedin Campus for these malformations will and 2Department of Medicine, Christ- nity-dwelling older adults, of sharpen diagnosis and prognos- church Campus, University of Otago, which 12,984 (17.2%) were tication for patients. However, New Zealand, 3Department of Phar- diagnosed with dementia. Of the heterogeneous nature of macy and Pharmacology, University these, 49.5% (6,430) individuals these abnormalities creates of Bath, Bath, United Kingdom. were prescribed at least one challenges for assigning patho- A high prevalence of MAP. By incorporation of the genicity. The aim of this study prescribing potentially inappro- recommendations, we observed was to apply high throughput priate medications, particularly a mean reduction of the ACB sequencing methods to a cohort medications with anticholin- by 0.49 (95% CI, 0.47-0.51). of 205 patients with periven- ergic properties (MAP), has been We could theoretically reduce tricular nodular heterotopia observed in older adults (>65 the prescription of MAP by (PVNH), a cortical malformation years of age) with dementia in suggesting therapeutic alterna- characterised by grey matter New Zealand. MAP have been tives in 2,006 older adults with nodules abutting the lateral associated with several adverse dementia (31.2%). ventricles of the brain due to a outcomes in this vulnerable failure in neural migration. population, specifically the risk The list of alternatives to MAP is intended to be a useful tool for of accelerating cognitive decline Aligned sequence data from health professionals that manage in patients with pre-existing patients was processed using individuals with dementia. The dementia. The purpose of this a range of custom pipelines implementation of the recom- study was to compile a compre- designed to capture short mendations for prescribing hensive list of therapeutic genetic variants, as well as large therapeutic alternatives to MAP alternatives to MAP prescribed structural events and somatic in this vulnerable population for comorbidities in individuals mutations. Data for parental along with an awareness created with dementia. Further, the list samples was available for 137 among prescribers has the was used in a cohort of patients patients, permitting filtering potential to reduce untoward with dementia, to determine based on inheritance models, effects associated with the theoretically if a reduction in along with identification ofde prescription of MAP. the anticholinergic burden novo events and causal biallelic (ACB) could be achieved. genotypes. A filtering method Building a brain – from based on recurrence with An extensive literature review known and suspected disease of ACB scales and serum anti- genes to phenotypes. genes was adopted for patients cholinergic activity of numerous 1,2 1 BJ Halliday , ZA Jenkins , DM without complete parental data. medications was conducted. Markie2, SP Robertson1. The list was applied to the 1Department of Women’s and Pathogenic variants were individuals who had a stan- Children’s Health, Otago Medi- identified for 17 patients with dardised comprehensive clinical cal School – Dunedin Campus, parental data, along with assessment, the International University of Otago, Dunedin, several candidate variants of Resident Assessment Instrument

uncertain diagnostic signifi- tered (P.O.) vehicle, metformin Brain function depends on cance. Of these variants, only (100 mg/kg), crizotinib (25 mg/ a balance between excitation two genes were recurrently kg) or the combination once and inhibition. Feed-forward mutated, FLNA (n = 4, P < 0.001, daily for 14 days. On the 15th inhibition (FFI) within the binomial) a known PVNH day serum was extracted for brain controls the firing of gene, and SON (n = 5, P < 0.001, ALT and creatinine analysis. principal excitatory neurons binomial), associated with the For efficacy testing, Nu/J mice and prevents runaway exci- multisystem developmental (n = 6) were subcutaneously tation. It is primarily mediated disorder ZTTK syndrome. injected with ALK+ H3122 cells by parvalbumin-expressing Using known and candidate in the flank region. Mice were (PV+) inhibitory interneurons. disease-gene patterns, another administered the same dosing Abnormal functioning of 12 pathogenic variants were regimen as in the toxicity study. these interneurons leads found in patients without Tumor volumes were weighed to neurological disorders, parental data, including an daily and full necropsies were including epileptic seizures. additional SON variant. performed on day 15. In the cortico-thalamocortical (CTC) network, dysfunctional The identification of six The serum markers of liver FFI has been implicated in patients with SON variants (ALT) and kidney (creatinine) absence seizure generation. The heavily implicates it in toxicity both remained under hallmark of absence seizures the pathogenesis of PVNH, the normal threshold for is spike-wave discharges increasing our understanding of all treatment groups. When (SWDs) measuring 3-4 Hz on an the molecular pathways critical examining the efficacy of the electroencephalogram (EEG) for neurogenesis. Further, the drugs; metformin, crizotinib with concomitant behavioural association between PVNH and and the combination signifi- arrests termed absences. Our ZTTK syndrome is underap- cantly decreased tumor volume laboratory has previously preciated and may provide a compared to vehicle (612, 424 reported defects in the acti- distinctive radiological marker and 552 vs. 943 mm3, respec- vation of PV+ interneurons in pointing to this diagnosis in the tively, P < 0.001). However, the the stargazer mouse model of absence of genetic testing. combination produced no added absence epilepsy, which could benefit when compared to crizo- Supported by a University of underlie hypersynchronous tinib alone. Otago Doctoral Scholarship. excitation leading to seizures. The combination produced The aim of the current study Repurposing the no additional toxicity and while was to investigate the impact of dysfunctional FFI within CTC hypoglycemic agent, metformin alone was able to slow tumor growth, the combi- network on absence seizure metformin, for nation did not have any greater generation and behaviour. targeted lung cancer. therapeutic benefit compared We used Designer Receptors to the monotherapies. We AR Bland, N Shrestha, RL Bower, Exclusively Activated by hypothesize that crizotinib is RJ Rosengren, JC Ashton. Designer Drug (DREADDs) limiting the entry of metformin technology to silence/excite PV+ Department of Pharmacology into cells via an inhibition on interneurons. To target these and Toxicology, School of organic cation transporters. Biomedical Sciences, Univer- interneurons, mice expressing Nevertheless, metformin sity of Otago, Dunedin. Cre recombinase in PV+ inter- alone produced a significant neurons (PV-Cre) were bred Lung cancer accounts for difference in tumor growth with inhibitory Gi-DREADDs the highest incidence of cancer compared to the control. This or excitatory Gq-DREADDs mortality worldwide. The provides justification to further mice. We confirmed selective EML4-ALK chromosomal rear- examine the effect and mecha- expression of DREADDs in rangement is involved in 2-7% nisms of metformin in cancer. of lung cancer cases. Crizotinib, PV+ interneurons via confocal the first-line treatment for Supported by a University of microscopy. Simultaneous ALK+ lung cancer is an effective Otago Doctoral Scholarship. video/EEG recordings were treatment, however, resis- made after silencing/exciting tance develops usually after An investigation into PV+ interneurons within CTC microcircuits. one year. To prevent/overcome dysfunctional feed- resistance, novel strategies are Silencing PV+ interneurons being explored. Epidemiological forward inhibition within the cortico- generated absence-like SWDs studies show a reduction in the and reduced ambulation in risk of cancer with the use of a thalamocortical Gi-DREADD animals. The mean hypoglycemic agent, metformin. network on absence duration and frequency of We aimed to test a combination seizure generation. SWDs were 2.9 ±0.3 sec and of metformin and crizotinib for 5.5 ±0.5 Hz, respectively. SWDs toxicity and efficacy in a xeno- S Panthi, B Leitch. were blocked by administrating graft model of lung cancer. Department of Anatomy, the anti-absence epileptic drug School of Biomedical Sciences, For toxicity analysis, balb/c ethosuximide. Conversely, Brain Health Research Centre, activating PV+ interneurons mice (n = 6) were adminis- University of Otago, Dunedin.

during pentylenetetrazole (PTZ) effective in many CRC patients. T Zeller2, PP Jones1, RR Lamberts1. induced seizures in Gq-DREADD The aims of this study were to 1Department of Physiology, animals delayed the latency, determine iR expression on HeartOtago, School of Biomedical decreased mean duration and T cells in CRC and associate Sciences, University of Otago, total number of PTZ-induced iR expression on T cells with Dunedin, 2University Heart and absence-SWDs. T cell functionality ex vivo Vascular Centre, University and in vitro. We hypothesised Medical Centre Hamburg-Eppendorf, These data indicate that loss Hamburg, Germany, 3Department that expression of iRs does of FFI within the CTC network of Surgical Sciences, Otago Medical not correlate with impaired is one causative mechanism for School – Dunedin Campus, Dunedin immune cell function and iR+ T 4 pathological SWD oscillations. Hospital, Dunedin, Department cells are important in the anti- This could be a target for future of Medicine, HeartOtago, Otago tumour immune response. Medical School – Dunedin Campus, improved treatment strategies University of Otago, Dunedin. since current anti-epileptic High dimensional analysis drugs are ineffective or cause of mass cytometry data from Long-chain acylcarni- serious side-effects in one-third patient samples (n = 8) identified tines (LCACs) are metabolites of patients. heterogeneous populations of essential for lipid metabolism iR+ T cells. Two populations in the heart. Recent metabo- Supported by a University of of T cells were significantly lomic studies have revealed that Otago Doctoral Scholarship and enriched in the tumour high plasma levels of LCACs, a Roche Hanns Möhler Doctoral compared to blood of CRC especially the 18:1 species, are Scholarship. patients: CD4+FOXP3+BLIMP-1+ associated with increased risk “effector Tregs” (P = 0.0009, of cardiovascular diseases, Inhibitory receptor Dunn’s multiple comparisons); including atrial arrhythmias. expressing T cells and CD4+PD-1+CTLA-4+CD39+ T This study aimed to address cells (P = 0.035, Dunn’s multiple whether LCAC 18:1 directly produce functional alters the susceptibility for cytokines and are comparisons). These iR+ T cells, enriched in the tumour, arrhythmias and contractility of enriched in the secreted anti-tumour molecules human heart muscle. tumour compared (Granzyme B, IFNg, TNF; n = 8). Human heart muscles (N = to the non-tumour Therefore, iR expression does 32) were mounted in a tissue bowel and peripheral not indicate impaired function superfusion bath and stim- blood of patients with within CRC tumours. PD-L1 ulated to contract were at 1 ligation in vitro did not signifi- colorectal cancer. Hz. The propensity for spon- cantly alter cytokine expression taneous muscle contractions J Harte1, J Leman1, F Munro2, (IL-2) or proliferation (Ki67) in in the absence of external J McCall2, H McGuire3,4,5, R Kemp1. PD-1+CD4+ T cells. However, stimulation was used to PD-1 ligation decreased cytokine 1Department of Microbiology and assess ex vivo arrhythmias. Immunology, Otago School of Medi- production and proliferation Relative to baseline conditions, cal Sciences, University of Otago, in CD4+FOXP3+ T cells (n = 4, exposure to LCAC 18:1 at a 25 Dunedin, 2Department of Surgical P = 0.0625, Wilcoxon sign-rank µM dose for 45-minutes (n = Sciences, Otago Medical School test), suggesting these two T cell 8) increased the proportion of – Dunedin Campus, University of populations respond differently spontaneously active muscles 3 Otago, Dunedin, Ramaciotti Facility to PD1 signalling. Ongoing in by 50% (Fishers exact test, P for Human Systems Biology, The vitro tumour conditioned media < 0.05). Furthermore, LCAC University of Sydney and Cente- experiments will determine the nary Institute, Sydney, Australia, 18:1 concurrently increased effect of the tumour microenvi- 4Discipline of Pathology, School the contraction force of the of Medical Sciences, Faculty of ronment on these populations. muscles in a dose-dependent Medicine and Health, The Univer- This study identifies popula- manner (1, 5, 10, and 25 µM, sity of Sydney, Sydney, Australia, tions of functional iR+ T cells n = 8 for each), with the 25 5Charles Perkins Centre, University in CRC tumours, which may µM dose inducing a 1.5-fold of Sydney, Sydney, Australia. explain a lack of clinical efficacy increase (P < 0.01, repeated Expression of inhibitory of ICB in these patients. measures one-way ANOVA). receptors (iRs), such as CTLA-4 Both the arrhythmogenic and and PD-1, are associated with Supported by a PhD schol- inotropic effects of LCAC 18:1 impaired immune cell function arship from the Cancer Society of were reversed following LCAC and can limit anti-tumour New Zealand. wash-out. To gain mechanistic immune responses. Therapies insight, recombinant HEK293 that block inhibitory receptors Arrhythmogenic and cells were loaded with the 2+ (immune checkpoint blockade, inotropic effects of long- cytosolic Ca indicator, fluo-4, ICB) can enhance the anti- and were superfused with tumour immune response and chain acylcarnitines 25 µM LCAC 18:1. LCAC 18:1 improve patient survival. A in the human heart. induced a marked cytosolic Ca2+ high infiltrate of iR+ T cells HM Aitken-Buck1, I van Hout1, J overload in this cardiac cell are present in the tumours of Krause2, PJ Davis3, RW Bunton3, DJ model, as well as a reduction in patients with colorectal cancer 2+ Parry3, MJA Williams4, S Coffey4, the intracellular Ca store size. (CRC); however, ICB is not These effects were augmented

by increasing extracellular Ca2+ arterial wall are responsible for of CaMKII (δ and γ). Results concentrations, suggesting that life threatening events including showed calcium calmodulin-de- LCAC 18:1 enhances Ca2+ flux stroke. The vascular wall is pendent protein kinase II across external and internal exposed to a number of stresses (CaMKIIδ) as the predominant membranes. that require cellular responses isoform. We next employed a to maintain homeostasis. The genetic approach by crossing This study is the first to show calcium/calmodulin-dependent ApoE-/- mice with CaMKIIδ-/- to that LCAC 18:1 can promote protein kinase II (CaMKII) generate an ApoE-/-CaMKIIδ-/- arrhythmias and contractility isoform family has important (dKO) mouse model. Analysis changes in human cardiac roles in maintaining vascular of the aortic sinus at 20 weeks muscle, which are linked to homeostasis but more recently showed extensive athero- cytosolic Ca2+ overload. This first emerged in the context of sclerosis in female groups. use of human heart tissue is a vascular dysfunction. Female dKO mice had a 162 critical step in improving the ± 31 µm3 reduction in aortic translatability of LCAC patho- Previously, we showed that sinus lesion volume compared physiology and metabolomics to systemic inhibition of CaMKII to female ApoE-/- (mean ± SEM, a clinical setting. leads to a reduction of athero- n = 11, P < 0.05, t-test). Finally, sclerosis in the brachiocephalic Supported by a University of adeno-associated viral vectors artery of a mouse model of Otago Doctoral Scholarship. (AAVs) expressing either atherosclerosis (ApoE-/-). The CaMKIIδ-mCherry or control- next critical step in translating mCherry were introduced to Calcium calmodulin this work to a clinical inter- ligated carotid arteries of dKO vention is to investigate which dependent protein mice and 4-weeks later the isoform of CaMKII contributes kinase II δ in the extent of lesion development to atherogenesis. pathogenesis of was assessed to show the atherosclerosis. The aortic tree of 13- and contribution of CaMKIIδ in 20-week ApoE-/- mice was atherogenesis. LPI Worthington, JR Erickson, dissected and the aortic arch Collectively, our results AK Heather. and carotid artery isolated. In show that CaMKIIδ is a major Department of Physiology, addition, human umbilical vein promoter of atherosclerosis School of Biomedical Sciences, endothelial cells and human lesion development. Impor- University of Otago, Dunedin. coronary artery smooth muscle tantly, we have identified a Atherosclerosis is the leading cells were cultured. PCR and specific therapeutic target that cause of death in the developed Western blotting was carried could provide impetus for drug world. The build-up of athero- out in all samples for the two development. sclerotic lesions within the primary vascular isoforms

URL: www.nzma.org.nz/journal-articles/abstracts-for-the-255th-otago-medical-school-research-society-phd-student-speaker-awards-wednesday-19-august-2020

Proceedings for the 254th Otago Medical School Research Society Research Speaker Awards, Wednesday 24 June 2020

Network meta-analysis priate (aerobic and anaerobic) The collagen mesh found in of antibiotics and bowel cover when combining IV and the extracellular matrix (ECM) OA, and those with appropriate of the myocardium plays a preparation in elective IV antibiotic cover. Sub-analyses critical role in preserving tissue colorectal surgery were performed using studies architecture and mechanical K Clifford, JC Woodfield, B Schmidt, examining rectal, left-sided, and properties. Dynamic remodelling G Turner, MA Amer, J McCall. right-sided regions of the colon. of the ECM occurs in different myocardial pathologies; this can Department of Surgical Sciences, We identified 75 RCTs manifest as a change in chamber University of Otago, Dunedin. including 16,891 patients. Treat- geometry and/or mechanical ments compared MBP+IV (5,642 There are discrepancies in properties which ultimately patients), IV (2803 patients), guidelines on bowel preparation impacts overall cardiac perfor- IV+E (397 patients), IV+OA± for colorectal surgery. While mance. Therefore, precise E (649 patients), MBP+IV+OA intravenous (IV) antibiotics quantification of collagen in the (4821 patients), MBP+OA (2093 are commonly administered, myocardium is of particular patients) and OA (486 patients). the use of mechanical bowel importance in histological The likelihood of SSI was signifi- preparation (MBP), enema (E) characterisation of myocardial cantly lower for IV+OA±E (rank and/or oral antibiotics (OA) is diseases. controversial. This controversy 1) and MBP+IV+OA (rank 2) stems in part from the historical when compared to other treat- Using Masson’s trichrome use of inadequate IV antibiotics. ments (IV, MBP+IV, MPB+OA and (MTC) to detect collagen is Our aim was to summarise OA), both P < 0.001. The addition a widely accepted standard all data from randomised of OA to IV antibiotics reduced in histopathology. However, controlled trials (RCT) by using SSI by approximately 50%. There manual examination of network meta-analysis (NMA) were minimal differences in AL collagen in stained slides can to determine the ranking of and in secondary endpoints. be time-consuming and is non-quantitative, while off-the- different bowel preparation This NMA supports the shelf computational approaches treatment strategies. This NMA addition of OA to IV antibiotics lack scalability and have is the first comprehensive for patients undergoing elective rarely been formally validated. review of this topic that colorectal surgery. Additional Accordingly, we developed accounts for the impact of anti- research should assess the a highly scalable and repro- biotic type and compares the effectiveness of IV+OA±E and ducible Machine Learning (ML) efficacy of all bowel preparation MBP+IV+OA. options in reducing postoper- pipeline to quantify the amount ative infections. Supported by a University of of collagen in MTC-stained Otago Research Grant. Whole Slide Images (WSIs). NMA was performed according to PRISMA guidelines. The ML pipeline consists RCT of adult patients under- A scalable, efficient of three distinct phases. 1) going elective colorectal surgery machine learning Preprocessing—splits the cover were included. The search pipeline to quantify WSI into smaller images included Medline, Embase, myocardial collagen (tiles) and removes any tiles with no tissue present. 2) ML Cochrane and SCOPUS data- in whole slide bases. Primary outcomes were inference—applies the ML wound infection (SSI) and anas- histological images. algorithms, namely, K-Nearest Neighbors (K-NN), Support tomotic leak (AL). The NMA was L Ariyasinghe, M Moharram, S Coffey. performed in Stata v15.1, using Vector Machines (SVM) and Department of Medicine, Otago three models: all identified Random Forest (RF). The ML Medical School, Dunedin Campus, algorithms were trained on 4567 studies, studies with appro- University of Otago, Dunedin.

pixels from manually annotated of Otago, 2Department of Pathol- hydrolases, designated fluo- tiles of right atrial appendage ogy, Stanford University, USA, rophosphonate-binding 3 samples, obtained from patients Department of Medical Biology, hydrolases (Fphs), which undergoing cardiac surgery. 3) The Arctic University of Norway, contribute to virulence of S. Norway, 4Shanghai Institute Segmentation—creates a fully aureus in the biofilm niche. of Materia Medica, Chinese segmented tile by classifying Academy of Sciences, China Here we report a struc- each pixel into one of the three ture-function characterization classes, specifically, collagen, Around a third of healthy of one of these serine hydro- non-collagen and blank. The humans are carriers of Staphy- lases, FphF, expressed during results from each tile are then lococcus aureus, they have the biofilm forming conditions. aggregated to provide a value bacteria on their skin without We determined that FphF is for the WSI as a whole. any active infection or disease. a promiscuous enzyme, able Despite being harmless in most to cleave hydrophobic lipid We estimated the classifi- individuals, S. aureus can cause substrates with a range of acyl cation accuracy using a test pathogenic infections. It often chain lengths. Using this newly set of 87200 pixels, obtained exists in biofilms in human acquired structural data and from samples from 10 patients. tissue, resulting in a biomo- similarities among the Fph Consequently, reported lecular matrix that is largely family, we show that other Fph accuracies are as follows: impermeable to the immune proteins, including FphB which K-NN—98.3%, SVM—99.6% and system and many traditional was linked directly to virulence, RF—98%. In addition to this antibiotics. The increased occur- may have a more well-de- remarkable accuracy, our ML rence of community-acquired fined substrate specificity. Our pipeline provides workflow effi- antibiotic-resistant S. aureus structural and biochemical ciency and ability to scale up to strains, often linked to biofilm studies confirm that FphF is handle large number of WSIs. formation, is a major health distinct from previously char- threat, requiring urgent devel- Supported by funding from the acterized enzymes, making it opment of new diagnostic and Department of Medicine and the an important reference enzyme therapy options. New Zealand Heart Foundation. in the serine hydrolase super- Serine hydrolases are a large family. Overall, our results Structural basis family of enzymes that play provide the first insight into the specificity and the mechanism for active-site key roles in bacterial homeostasis and survival at the of substrate and chemical probe probes targeting host-pathogen interface during binding to the Fph protein Staphylococcus aureus infection. They play a role in family. This information will serine hydrolase biofilms, contributing to the aid in future efforts to targeting virulence factors. difficulty of achieving effective serine hydrolase virulence treatment. This makes serine factors from S. aureus and other M Fellner1, CS Lentz2,3, SA Jamieson1, hydrolases promising new related bacteria. JL Brewster1, L Chen2,4, M Bogyo2, anti-virulence and anti-infec- This work was supported PD Mace1. tivity targets. by a University of Otago 1 Biochemistry Department, School Health Sciences Postdoctoral of Biomedical Sciences, University Activity-based profiling identified a family of serine Fellowship.

URL: www.nzma.org.nz/journal-articles/proceedings-for-the-254th-otago-medical-school-research- society-research-speaker-awards-wednesday-24-june-2020

Abstracts for the 253rd Otago Medical School Research Society Summer Student Sepeaker Awards, Wednesday 22 April 2020

The developmental of Sox2 and Tubb2b, markers Dunedin School of Medicine, University of Otago, Dunedin. effects of disrupting of neuronal proliferation, were examined by immunohisto- DONSON – a gene Dementia, characterised chemistry and qPCR. by progressive deterioration involved in genetic of cognitive functions, is a There was a 55% decrease microcephaly. significant and growing health in Sox2 and a 69% decrease challenge, with a huge social MR Mulligan1, DE Jenkins1, in Tubb2b mRNA (markers of and economic impact. Current E Damsteegt2, C Beck2, LS Bicknell1. neural proliferation) in anterior available treatments have 1 tissues in the Donson 800 pg Department of Pathology, limited efficacy, warranting new Dunedin School of Medicine, edited tadpoles relative to the targeted interventions. Several 2Department of Zoology, Univer- controls. H&E staining alongside studies demonstrate dysfunc- sity of Otago, Dunedin. markers of neuronal prolifer- tional electrical activity in ation, suggested a reduction in The clinical features of many brain-wide functional networks, the area of the fourth ventricle genetic disorders are often particularly involving the and decreased thickness and caused by complex mechanisms dorsal anterior cingulate cortex organisation of the developing that are poorly understood. (dACC), in individuals with brain, specifically in the Donson Recently, mutations in DONSON, cognitive impairments. Non-in- 800 pg tadpoles, relative to the a DNA replication protein, vasive transcranial electrical controls. have been found to cause stimulation, can normalize microcephaly (reduced brain The morphological changes dysfunctional brain activity, size) in children. The develop- suggested alterations to prolifer- thereby improve cognition. This mental processes that cause ation, which would fit with the pilot, double-blinded (partici- this reduction in size remain role of DONSON in DNA repli- pants and assessor) randomised unknown. This project aimed to cation. This is further supported sham-controlled parallel study the effects of disrupting by a reduction in transcripts trial evaluated the safety and Donson using the CRISPR Cas9 for neural proliferation. explored the immediate trend of editing tool in the developing Further studies will provide effect of a novel, dACC targeted, brain of Xenopus laevis (African further insight into the exact high-definition transcranial clawed frog). mechanism underlying such a infraslow pink noise stimu- To study how disruption of reduction in proliferation, to lation technique (HD-tIPNS) Donson affects the developing understand the link between on cognition in healthy older brain, Xenopus embryos were DONSON and microcephaly. adults. microinjected with Donson Supported by a University of Ten cognitively healthy partic- sgRNA at 400 or 800 pg/embryo Otago Division of Health Sciences ipants were randomised to plus Cas9 protein, with control Summer Research Scholarship. receive either HD-tIPNS or Sham embryos uninjected or injected stimulation, for a single session with only Cas9 protein. PCR of High definition of 30 minutes. Adverse effects the cleavage site and T7 endo- were recorded throughout the nuclease was used to confirm transcranial infraslow intervention period. Cognitive editing status in DNA from pink noise stimulation tests [Stroop tests (ST), trail injected tadpoles. Four tadpole for improving executive making test (TMT), and digit heads from each condition functioning in healthy span recall test (DSRT)] and were fixed and sectioned into 4 older adults – A resting-state electroencepha- µm sections, which underwent lography (EEG) were conducted haematoxylin and eosin (H&E) pilot safety trial at baseline and immediately staining to analyse general F Whittington, D De Ridder, D Adhia. post-intervention. Descriptive histology. Expression profiles Department of Surgical Sciences, statistics were used to explore

the trend of effect of HD-tIPNS Participants received one 1Department of Preventive and Social on cognitive outcomes. EEG data of the following medications Medicine, Dunedin School of Medi- were analysed using sLORETA (0.5 mg/kg or 1.5 mg/kg oral cine, University of Otago, Dunedin. to explore changes in brain ketamine or 0.02mg/kg oral The National Coronial activity (BA) and functional midazolam (psychoactive Information System (NCIS) is a connectivity (FC). control)). One hour later, they repository containing inves- experienced a 5-level VR spider tigative data of all deaths No adverse outcomes were simulation. At higher simu- reported to a Coroner in reported. All ST variables lation levels, spiders numbers Australasia. A Coronial investi- improved post-treatment, increased, as did their size and gation can include toxicological irrespective of the group. TMT movement. We obtained vital analysis at the discretion of demonstrated a positive trend signs measurements and a the Coroner. Little is known of effect in the HD-tIPNS group. visual analogue scale (VAS) for both within New Zealand, and No effect was observed in DSRT. anxiety (range 0–100) before internationally, about potential A trend towards significant and at each level of the VR biases related to toxicology effects of HD-tIPNS was evident spider simulation. Participants report requests. on the infraslow, alpha, and returned for all three sessions. theta bands in dACC network (P This study undertook a retro- The study is ongoing, with data = 0.06), compared to sham (P = spective review of New Zealand presented for 11 participants. 0.93). Increased and decreased injury deaths in 2014. Data FC of dACC with the dorso- Significantly more simulation was collected from the NCIS lateral prefrontal cortex and levels were completed after and the Mortality Collection. the posterior cingulate cortex 1.5 mg/kg ketamine compared Variables of interest, such as respectively, was observed in with midazolam dosing (mean circumstances of death and HD-tIPNS group (P < 0.001). (SD) 1.0 (1.2), t = 2.6, P = 0.03). characteristics of decedents, A two-way repeated measures were extracted from the These results implicate the analysis of variance identified Coronial files. Eligible injury-re- safety and ability of HD-tIPNS significant treatment x time lated deaths referred to the technique to modulate the interactions for VAS anxiety and Coroner that had toxicological activity and FC of dACC, respiration rate (RR) during the investigations were compared suggesting that it has potential simulation (VAS anxiety F = 2.02, to those that did not. for use as a treatment tool. df = 8, P = 0.05; RR F = 2.94, df=8, Of the 744 unintentional Supported by a scholarship from P = 0.007). Heart rate changes injury Coronial cases in those the Australia and New Zealand were not statistically significant under 85 years of age, 560 (75%) Society of Geriatric Medicine. between dose groups. received toxicological analysis. Most of the assessments Females were less likely to Pilot study of reported support our hypothesis have a toxicology report (67%; ketamine in phobic that ketamine has an anti- males 78%; 95%CI for difference participants using phobic effect in patients with -5%, -18%), as were the young virtual reality stimuli. spider phobia, compared with and the elderly (58% and 55% midazolam. This is consistent respectively compared to 88% A Krijgsman, S.Gallagher, with our earlier research and for those aged 25–34). Māori S Neehoff, P Glue. supports a theory that ketamine were more likely to recieve Department of Psychological Medi- may work across a broad range toxicological analysis compared cine, Dunedin School of Medicine, of internalizing disorders via to NZ European (83% and University of Otago, Dunedin. effects on a central neuroticism 73% respectively; 95%CI for Treatment for specific process. This study adds support difference 3%, 17%). Only 44% phobia, an anxiety disorder to the use of ketamine as an of decedents that died as a characterized by unreasonable acute treatment for specific result of a fall received a toxi- fears related to specific objects phobia. cology report. There was a clear relationship between receiving or situations, is limited to Supported by the Division of a toxicology report and the time psychological interven- Health Sciences. tions with no approved drug between injury and death; 86% treatments. Recent research for those that died on the same identified an incidental effect Factors related day as their injury compared of low-dose injected ketamine to the presence to 13% in those that this period on anxiety ratings for a specific of toxicological was over a week. Opioids were phobia (blood/needle phobia). investigation detected in 23% of cases that received toxicological screening This study aimed to extend this following fatal injury: finding by using oral ketamine tests. (minimal side effects), a more a retrospective quantitative review It is currently unknown common phobia (spider phobia), whether the observed patterns and a more controllable experi- of Coronial records are justifiable or whether they mental paradigm (virtual reality in New Zealand. relate to subconscious biases. (VR) exposure), in a cross-over Follow-up qualitative research L Nie1, G Davie1, R Lilley1. active-controlled design. is required to understand this

further. Better understanding of subsequent inverse agonism. predictor to evaluate the cancer Coronial processes helps inform Radioligand binding compe- stage of patients. The objective policymakers, researchers, and tition assays performed of this study was to develop a practitioners, and contributes to demonstrated that ABM300 regression model that updates injury prevention and improved has an enhanced ability (pEC50 the calculator with serum thyro- health policies. 7.133 ± 0.201) compared to globulin information. ORG27569 (pEC50 6.316 ± 0.101) Supported by a Otago In this study, we used both in increasing the binding of University Division of Health the logistic regression and CP55,940 to the CB1. We also Sciences Summer Studentship. the LASSO methods for model found that ABM300 (10 µM) building. Records from 3962 abolishes the effect of CP55,940 thyroid patients were analysed ABM300, a new negative on phosphorylation of ERK1/2 for training models for recur- allosteric modulator (PerkinElmer AlphaLISA rence prediction. A-priori Surefire kit), resulting in of the CB cannabinoid twelve variables were inves- 1 pERK1/2 levels at 23% ± 6.8 of receptor, exhibits a tigated (age at operation, maximum CP55,940-induced sex, number of carcinomas similar mechanism of ERK1/2 phosphorylation, similar presented in the operation, action as ORG27569. to ORG27569. Furthermore, size of the greatest tumour, pilot data indicated that both N Khalajiassadi, D Finlay, histologic type of carcinoma, ABM300 and ORG2769 lead M Patel, M Glass. extrathyroidal extension to a concentration-dependent status of tumours, pathologic Department of Pharmacology decrease in CP55,940-mediated and Toxicology, School of staging of the primary tumour, β2-arrestin recruitment. Biomedical Sciences, Univer- presence of venous invasion of sity of Otago, Dunedin. This study has made a start the primary tumour, immuno- The extensive role of the in understanding the ABM300 histochemistry for the primary endocannabinoid system in mechanism of action and found tumour, presence of extranodal homeostatic regulation and strong similarities to that of spread, number of lymph nodes the fact that G protein-coupled ORG27569; that is, that the and serum thyroglobulin level receptors (GPCRs) are the most allosteric modulator prevents presented in the scans) as common drug targets makes the activity of the orthosteric predictors of recurrence. the cannabinoid receptor 1 agonists. Both approaches demon- (CB1) a suitable subject for Supported by a Fastier strated excellent performance. development of novel thera- Summer Student Research Fund Logistic regression (with an peutics. Recently, research on scholarship. AUC of 0.874) had a slightly CB1-targeting drugs has shifted better performance, whereas to allosteric modulators in lasso regression (with an AUC the hope that they offer fewer Thyroid cancer of 0.856) utilized fewer vari- adverse effects than orthosteric recurrence prediction ables to achieve a similar result. ligands. Allosteric binding sites using regression The LASSO method had iden- are distinct from orthosteric approaches. tified five crucial variables for sites and are less conserved in predicting structural recurrence comparison, allowing allosteric J Li1, J Zeng1, R Turner2, of thyroid cancer. These include ligands to exhibit a greater C Nylén,3 M Sywak3. extrathyroidal extension status receptor selectivity. In contrast 1 Department of Preventive and of tumors, the pathologic staging to promising in vitro studies, Social Medicine, Dunedin School of the primary tumor, the the existing allosteric modu- of Medicine, 2Centre for Biostatis- presence of extranodal spread, lators such as ORG27569 have tics, Division of Health Sciences, the number of lymph nodes that not demonstrated substantial University of Otago, Dunedin, 3 are involved by carcinoma and activity in vivo. ABM300, a novel Endocrine Surgery Unit, University the serum thyroglobulin level allosteric modulator has been of Sydney, Sydney, Australia. presented in the scans. shown to alleviate abnormal Prediction of thyroid cancer behaviours in hyperdopami- recurrence is important Recurrence of thyroid nergic mice models. We aimed information for planning cancer can be predicted with to characterise ABM300 in vitro post-surgical follow-ups. Cancer reasonable accuracy with as (in HEK293 cells), in comparison stage is a well-known predictor little as five clinical variables. with ORG27569. of thyroid cancer and the This information can be used American Thyroid Association to help plan post-surgical Pilot data (n = 3) suggested offers a thyroid cancer staging follow-up of patients. Future that ORG27569 and ABM300 calculator that helps predict work will investigate whether exert similar effects on CB1 the stage of thyroid cancer these variables also predict cAMP signalling, as the co-ad- using preoperative clinical and survival. ministration of either allosteric demographical information. The modulator with potent CB1 current calculator, however, is Supported by a Dunedin School agonist CP55,940 leads to missing information on serum of Medicine Summer scholarship. blockade of agonism, and thyroglobulin which is a critical

Upregulation of GAD65 blotting detected significantly studies examining patients in somatosensory increased GAD65 in somato- with symptomatic heart failure sensory cortex of epileptic or rheumatic heart disease cortex of stargazer mice compared to non-epi- suggest a relationship between absence epileptic mice. leptic littermates (P = 0.0324, increased LAP and fractional N Lyons, S Panthi, B Leitch. Mann-Whitney U test, N = 11 exhaled nitric oxide (FeNO), Epileptic/13 Non-Epileptic). No which can be measured by a Department of Anatomy, significant change in GAT3 or reliable portable hand-held School of Biomedical Sciences, University of Otago, Dunedin. GAD67 levels was detected (P > analyzer. If such a relationship 0.05, Mann-Whitney U test, N = existed, a portable diagnostic Childhood absence 9 Epileptic/14 Non-Epileptic or test could be used to supplement epilepsy is one of the most 10 Epileptic/14 Non-Epileptic current assessments. common paediatric epilepsies. respectively); although GAD67 It is characterised This observational study trended towards elevation by frequent non-convulsive examined a subset of the in epileptic animals above seizures causing brief loss of OxVALVE cohort, a UK popula- controls. consciousness. Absence seizures tion-based cohort aged 65 years arise within the cortico-thal- GAD65 upregulation may be and older. Participants were amo-cortical (CTC) network; a mechanism of compensating eligible for inclusion if they however, the precise molecular for reduced FFI, and account for did not have a previous diag- mechanisms are not fully the heightened somatosensory nosis of valvular heart disease, understood and are likely cortex GABA levels previ- and could provide informed multifactorial. This may account ously observed. This discovery consent. Each participant had for the variability in patients’ provides the foundation for echocardiography performed at responses to antiepileptic drugs investigating the cell-type spec- their local general practice. The (AEDs). Although AEDs greatly ificity of GAD65 elevation and E/e’ ratio, measured via trans- improve quality of life for interrogating its role in absence thoracic echocardiography, was most patients, one third cannot seizure genesis and mainte- used as a surrogate of LAP. A be effectively treated and nance over the developmental consecutive subset of 277 partic- there is no cure for epilepsy. timespan. GAD65 may itself ipants had FeNO measurement Understanding underlying constitute a therapeutic target attempted using a NIOX VERO molecular causes is imper- or may allow discovery of an electrochemical analyzer. ative. The stargazer mouse underlying mechanism to thera- Mean age of participants model of absence epilepsy has peutically target. was 73.5 ± 6.3 years, with 45% a mutation reducing excitatory Supported by the Otago female. Only 10 participants had input to feed-forward inhib- Medical Research Foundation - an NYHA class of III or more itory (FFI) interneurons; FFI EMM Haynes Charitable Trust (New York Heart Association prevents runaway excitation Summer Research Scholarship. classification of heart failure in networks. The stargazer patients). FeNO was successfully mutation specifically affects CTC measured in 227 participants, parvalbumin containing GABA Investigation of with a mean FeNO of 24.2 + 15.6 (γ-Aminobutyric acid) inter- exhaled nitric oxide ppb. Mean E/e’ was 11.5 ± 3.5. neurons, causing FFI deficits as a measure of left There were 58 participants with and altered GABA levels. This atrial pressure. high E/e’ (>14), and these were study aimed to investigate older than those with normal whether changes in GABA levels S Jones1, A Prothero2, S Myerson2,3, E/e’ (<8) (76.8 ± 7.4 years vs 72.0 are due to altered expression B Prendergast4, S Coffey1,3. ± 5.5 years, P < 0.001). No rela- of its production enzymes, 1 Department of Medicine - HeartO- tionship between FeNO and E/e’ glutamate decarboxylases tago, Dunedin School of Medicine, was seen on regression analysis (GADs 65&67), and/or transport University of Otago, Dunedin, (linear regression R2= 0.007). proteins, GABA transporters 2Oxford University Hospitals NHS 3 (GATs 1&3). Trust, Division of Cardiovascular These results illustrate Medicine, Radcliffe Department that FeNO is not an accurate Confocal immunohistochem- of Medicine, University of Oxford, predictor of elevated E/e’ in istry of sections double labelled 4Department of Cardiology, St predominantly asymptomatic for parvalbumin and GAD/ Thomas’ Hospital, England. patients in a general practice GAT showed no differences During left-sided heart setting. in cortical expression pattern failure, left atrial pressure between epileptic and non-ep- (LAP) increases, producing Supported by a Heart Foun- ileptic animals. However, pulmonary congestion. Previous dation Summer Scholarship. semi-quantitative western

URL: www.nzma.org.nz/journal-articles/abstracts-for-the-253rd-otago-medical-school-research-society-summer-student-sepeaker-awards-wednesday-22-april-2020

NZMJ 25 June 2021, Vol 134 No 1537 ISSN 1175-8716 © NZMA 118 www.nzma.org.nz/journal published by the New Zealand Medical Association

NZMJDigest http://www.nzma.org.nz/publications/nzmjdigest

The NZMA publishes the e-magazine NZMJDigest 10 times a year. It contains news and views from the profession and the NZMA, including the NZMA Chair's editorial, along with highlights from and links to the New Zealand Medical Journal. Click on the image above to view the latest issue. We welcome contributions from members and readers. To contribute to the NZMJDigest, please email [email protected]